Beruflich Dokumente
Kultur Dokumente
Departament de Qumica
Unitat de Qumica Analtica
Hg0(aq)
O3
Hg2+(aq)
Los estudios que recoge la presente memoria forman parte de los siguientes
proyectos de investigacin:
-
Desarrollo
de
nuevos
materiales
en
tcnicas
de
separacin.
metals
in
Contaminated
Groundwaters
and
Soils
Summary
1 Introduccin
11
1.2.1
El mercurio en la atmsfera
13
1.2.2
El mercurio en suelos
14
1.2.3
14
1.2.4
15
16
1.3.1
16
1.3.2
16
17
18
1.5.1
Fuentes puntuales
18
1.5.1.1
Fuentes de combustin
19
1.5.1.2
Fuentes de produccin
19
1.5.1.3
Fuentes diversas
20
1.5.2
Fuentes difusas
20
21
1.6.1
Cintica y metabolismo
22
1.6.2
Efectos en humanos
23
1.6.3
24
26
1.7.1
27
1.7.2
29
29
1.8.1
31
1.8.1.1
34
1.8.1.1.1
Extraccin
34
1.8.1.1.2
Separacin y determinacin
38
1.8.1.1.3
42
1.8.1.2
44
1.8.1.3
Desorcin trmica
46
1.8.1.4
Extraccin secuencial
47
49
1.9.1
50
1.9.2
52
1.10
53
1.10.1
54
1.10.2
55
1.10.3
56
1.10.3.1
57
1.10.4
58
1.10.5
59
1.10.5.1
60
1.10.5.2
61
1.11
62
1.11.1
Hidrxido de sodio,
62
1.11.1.1
Reactivos de partida
63
1.11.1.2
64
65
66
66
1.11.2
67
1.12
Objetivos
69
2 Metodologa y experimental
71
73
74
75
76
2.4.1
76
2.4.2
78
2.4.3
2.4.3.1
Stacking, FASS)
80
2.4.3.2
81
2.4.3.3
81
2.4.4
Aplicaciones
82
2.4.5
82
83
84
2.6.1
85
2.6.2
90
2.6.2.1
Dispersin (scattering)
91
2.6.2.2
Absorcin
91
2.6.3
92
2.6.4
95
2.6.5
96
2.6.6
97
2.6.7
98
separacin
100
2.7.2
100
2.7.3
Regeneracin de la resina
100
101
103
105
dopado
105
3.1.2
3.1.3
106
107
110
XANES
3.2.2
112
Aplicacin de XANES a la determinacin de especies de Hg en muestras
114
de suelos contaminados
3.2.4
112
114
116
119
117
122
122
123
4 Conclusiones
133
5 Anexos
139
Anexo 1
Stability study on a West-based methodology to determine organomercury
compounds in polluted soil samples; X.Gaona, M.Valiente; Analytica Chimica Acta
(2003), 480, 219-230
Anexo 2
XANES data evaluation of environmental samples: Comparison of WinXAS and
SpecManger programs; X.Gaona, S.Mangold, M.Valiente and W.Calmano, draft
Anexo 3
Direct speciation of mercury in polluted soils by XAFS techniques; X.Gaona,
S.Mangold, W.Calmano and M.Valiente, to be submitted
Anexo 4
Clean Ion-Exchange Technologies. 4. Modelisation and optimisation of the
conditions for a resin-based synthetic process of alkaline hydroxides; X.Gaona and
M.Valiente, to be submitted
Anexo 5
Summary
Summary
Summary
Summary
Summary
Mercury is a common element, which has several applications for humans not only in the
industry but also at home in our daily life. It is used since the ancient centuries and it has
been the basis for its own technological revolutions, like its introduction to thermometers.
In natural environments, it is found mainly as the very stable and insoluble form cinnabar,
in ores with a concentration varying from hundreds of parts per million to few percents of
the total weight. Almadn mines, the largest mercury mine around the world is located in
Spain, and it has been exploited since the Romans.
On the other hand, mercury is also a quite toxic element, and its uncontrolled presence
in the environment can become very dangerous for human health. As other pollutants, the
toxicity of mercury concerns with its chemical forms (chemical speciation). Interaction of
such speciation with the environment will also determine its mobility, a key parameter to
environmental impact. Therefore, the ability to recognise the different mercury species
has been identified as a milestone by the scientific community. Such ability allows for an
adequate classification of polluted environments as well as the assessment to identify most
appropriate remediation technologies.
Another way to fight for a pollutant reduction is to develop environmentally friendly
processes, i.e, avoiding toxic elements like mercury. For this reason, paints, batteries,
fungicides containing mercury have been substituted for other ones less toxic, whereas in
the industry, mercury-containing processes are being substituted (when possible) for harm
lesser ones.
Along this thesis and considering mercury as a global pollutant, both speciation from an
analytical point of view and mercury-less process development, have been studied.
With respect to chemical speciation, both polluted soils and underground waters have
been studied. High Performance Capillary Electrophoresis (HPCE) has been selected as
analytical separation technique for laboratory speciation (combined with the modified
West extraction method when dealing with soils or sediments). HPCE is becoming a very
appreciate tool for analytical determinations, due to its high separation capacity.
However, a very important drawback of the technique has been also pointed out: its poor
detection limits. This low sensitivity limits a direct application of the technique to
environmental problems, where normal concentrations are at ppb (even ppt) level. There
are mainly two solutions for such problem: the coupling with a high sensitive detection
techniques (i.e, ICP-MS) or the incorporation of a pre-concentration step in the process.
The first one use to be a good but expensive solution besides the problems with the
interface HPCE ICP-MS. Thus, the second option appears to be most appropriate since it
can be directly implemented to conventional equipments. Along this thesis, Large Volume
Summary
Sample Stacking (LVSS) without polarity switching was applied. In this case, a methodology
applied to proteins determination was used, consisting on an stacking effect achieved by
the pair acetonitrile sodium chloride. Due to the large amount of parameters to be
considered for the method optimisation, an Experimental Design was used. Therefore, by a
27-6IV design, the seven initial variables were reduced to five. Afterwards, a Face Centred
Cube design was used, achieving a model response which allowed to determine the optima
conditions for mercury analysis. By using such conditions, the detection limit was reduced
in one order or magnitude (100 ppb). However, this detection limit is still not low enough,
so further work will be required.
Considering mercury speciation in soils and sediments by conventional technologies
(that is extraction + chromatographic separation), the possible modification of the species
during the extraction process has been largely discussed and studied by the scientific
community. In this sense, a stability test along the extraction and analysis process was
performed. The experiments conclude that cystine, a degradation subproduct of cysteine,
could stimulate the conversion of organic to inorganic mercury when combined with light.
Since several extraction methodologies still include cysteine as a purification step, such
effect must be taken into account. However, it is also true that considering the normal
time scale in an extraction experiment, the real risk to alter mercury species by this
method is relatively low.
As an alternative to the separation of chemical species, the direct speciation of mercury
in aqueous samples was studied by ICP-AES. This study is specially relevant, since no
efficient methodologies exist for such direct speciation. The basis of the method is the
high variation in sensitivity that methylmercury experiences with the solution media. Thus,
the behaviour (considering the slope of calibration curve) of methylmercury in cysteine
media is very similar to the behaviour shown by inorganic mercury both in cysteine and
gold media (gold is a typical media in mercury analysis). However, when methylmercury is
analysed in gold media, the increase of the sensitivity is almost three times. Such effect
allows the easy determination of methyl and inorganic mercury mixtures, by consecutive
analysis of the sample in cysteine and gold media.
Again, the problem of the method is its high detection limits, so a complementary work
with pre-concentration techniques (like SPE cartridges) is needed.
Concerning direct speciation techniques for mercury, the main contribution of this
thesis has been the incorporation of direct techniques for soil analysis using the
Synchrotron light. XANES (X-ray Absorption Near Edge Spectroscopy) as a XAS technique (Xray Absorption Spectroscopy) has been successfully applied for the determination of
Summary
different forms of mercury in soil and sediment samples. XANES is a very special X-ray
absorption technique, since the source used for the X-ray generation is not a common Xray tube but a synchrotron facility. The X-rays generated there have very special
properties, i.e. its high brilliancy and continuity in the complete energy range. Such
properties provide very specific information on the elements present in the sample. The
strategy consists on irradiating the sample with an X radiation having energy close to the
energy of a deep core electronic level of the target element. The excited electron, when
interacting with the neighbouring atoms, provides useful information about the species
present in the sample.
In the present study, the basis of such analysis has been built up, establishing the
conditions at the corresponding beamline, the type of samples and their preparations, and
creating the know-how for the subsequent data treatment. In addition, the validity of the
method has been accomplished by correlation of XAS results with those obtained from
conventional speciation techniques (extraction + HPCE separation). It has been also
demonstrated the usefulness of the technique for qualitative data, i.e, determination of
the ionic character of mercury bonds.
In relation with the mentioned way to reduce pollutant impact, the second part of the
present studies has been dedicated to the development of an alternative technique for the
synthesis of alkaline hydroxides. It is well known that the process of mercury cathode for
the synthesis of sodium hydroxide is still in use in several countries, despite the existence
of some alternatives (membrane and diaphragm). Even if the regulation is becoming
stronger, and the measures for controlling releases more efficient, the fact is that mercury
cathode industries for sodium hydroxide production are a potential source of mercury
pollution.
Therefore, the developed methodology intended to be as friendly as possible with the
environment. For this purpose, ion-exchange resins technology was selected as the basis
for such process of synthesis using as reagents calcium oxide and the residue of sea water
desalination industries. A carboxylic resin was utilised, due to its high affinity toward
divalent ions. A complete set of experiments were done in order to find out the
appropriate thermodynamic and kinetic conditions for the synthetic process. Equally, a
complete set of experiments was performed for the optimisation of the regeneration
process (applying IXISS methodology).
The results were completed by an appropriate modelling of the process that considered
parameters such as the separation factor, the solubility product of calcium hydroxide or
the ratio m/V (mass of resin/volume of total dissolution). The influence of the particle size
Summary
of the calcium hydroxide present in the system was also studied since this parameter
appeared to strongly affect the solubility product of the compound (and in consequence
the final product composition)
Once completed, the resin-based process was compared with the mercury cathode
process, in terms of ecological impact. For this purpose, two parameters of the assessment
tool Life Cycle Assessment were applied. These parameters were: Global Warming and
Abiotic Resources Depletion. In addition, data about the use of uranium for energy
production and release of mercury during the process were taken into account. The direct
comparison leads to an unsatisfying result, since the resin process appeared to be more
pollutant in terms of Global Warming and Abiotic Resources Depletion. The explanation for
such unexpected result is found in the required final concentration of sodium hydroxide.
Thus, while for the mercury cathode process the final concentration is 50% w/w, the resinbased process achieves a maximum final concentration of around 0.6 M. Therefore, the
method drawback is the high amount of energy needed for evaporate the water, in order
to obtain a comparable final product in terms of concentration. Alternative methods for
solute concentration as well as the possibility to the direct use of the diluted sodium
hydroxide would dramatically decrease the environmental impact of the developed
method.
Introduccin
1 introduccin
Introduccin
Introduccin
Introduccin
finales del siglo XV haba sido escaso y casi exclusivamente como bermelln para la
fabricacin de pinturas, as como en medicina. Quien dio el primer impulso para el gran
consumo de mercurio fue el sevillano Bartolom Medina, al poner a punto en 1557 el
mtodo del Beneficio del Patio para la amalgamacin en fro de los minerales de plata. A
partir de ese momento las aplicaciones del mercurio empezaron a multiplicarse. As
Paracelso, en el siglo XVI, introdujo su empleo en el tratamiento de la sfilis; Torricelli, en
1643, lo utiliz en su barmetro; en 1720, Fahrenheit en su termmetro o a Priestley, a
quin sirvi de fundamento a finales del siglo XVIII para el anlisis de gases.
De entre sus propiedades fsico-qumicas, mencionar que el mercurio tiene un nmero
atmico de 80, un peso atmico de 200.59, una temperatura de fusin de 38.87C y una
temperatura de ebullicin de 356.58C. El mercurio elemental es un metal pesado, de
coloracin blanca-plateada, y lquido a temperatura y presin ambiente. A 20C su
gravedad especfica es 13.456 y la presin de vapor 0.16 Pa. As pues, la atmsfera
saturada a 20C contiene aproximadamente 15 mg/m3, valor 300 veces mayor que la
concentracin ocupacional (aire ambiente) mxima puntual permitida (0.05 mg/m3) o 1000
veces superior al nivel de exposicin continua (0.015 mg/m3)1. La mayora de mercurio que
se encuentra en la atmsfera es mercurio elemental en forma vapor.
El mercurio tiene siete istopos estables (196, 198, 199, 200, 201, 202, 204) y cuatro
istopos radioactivos inestables (194, 195, 197, 203). Puede existir en tres estados de
oxidacin: Hg0 (metalico), Hg22+ (mercurioso) y Hg2+ (mercrico), siendo sus propiedades
muy diferentes en funcin del mismo. Las formas mercricas y mercuriosas pueden formar
un gran nmero de compuestos orgnicos e inorgnicos, aunque la forma mercuriosa es
raramente estable bajo condiciones ambientales convencionales2,3.
La mayor parte del mercurio presente en el medio ambiente (exceptuando la atmsfera)
se encuentra en forma de sales mercricas inorgnicas y compuestos organomercricos.
Estos ltimos vienen definidos por la presencia de un enlace C-Hg covalente. As, los
compuestos de mercurio ms comunes en el medio ambiente son: las especies inorgnicas
de Hg(II) HgCl2, Hg(OH)2 y HgS; los compuestos de metilmercurio CH3HgCl y CH3HgOH; y, en
menores proporciones, otros organomercuriales como dimetilmercurio o fenilmercurio4.
Los compuestos de mercurio suelen permanecer en la fase acuosa como molculas no
disociadas, dando lugar a valores bajos de solubilidad. La mayora de organomercuriales no
1. Environmental Health Criteria 118: Inorganic Mercury, World Health Organization, Geneva, (1991)
2. DERA Report, D.S. Lee et al., Sources, Sinks and Levels of Atmospheric Mercury in the UK, (2000), [en lnea],
<http://www.aeat.com/netcen/airqual/reports/dera_hg/dera_hg1.pdf>, [5 Marzo 2004]
3. J.E. Fergusson, The Heavy Elements: Chemistry, Environmental Impact and Health Effects, Pergamon, Oxford, (1990)
4. EPA Mercury Study Report to Congress, Volume III, Fate and Transport of Mercury in the Environment, (1997), [en lnea],
<http://www.epa.gov/ttn/oarpg/t3/reports/volume3.pdf>, [5 Marzo 2004]
10
Introduccin
son solubles y no reaccionan con cidos dbiles o bases. El CH3HgOH es sin embargo muy
soluble debido a la gran capacidad del grupo hidroxilo para formar puentes de hidrgeno.
Por otro lado, las especies mercricas inorgnicas presentan una gran variedad de
solubilidades. HgCl2 es por ejemplo muy soluble en agua, mientras que HgS es
prcticamente inerte debido a la gran afinidad del mercurio por el azufre.
Para dar valores cuantitativos a dicho ciclo, es necesario conocer las cantidades de
mercurio presentes en los diferentes reservas terrestres. Se ha estimado que los ocanos
son, con diferencia, los mayores reservas de mercurio, con alrededor de 1017 g de mercurio
principalmente en forma de HgS. Las aguas ocenicas contendran alrededor de 1013 g, la
misma cantidad que suelos y sedimentos conjuntamente; la biosfera, principalmente en
forma de biota terrestre, contendra 1011 g; 108 g la atmsfera y 107 g las aguas terrestres.
En este balance quedara sin embargo excluido el mercurio presente en minas y otros
reservas subterrneos5-6.
Igualmente, es necesario aportar datos acerca de la presencia antropognica de
mercurio en el medio ambiente. Debido a la gran movilidad del mercurio dentro de su
ciclo, esta cuantificacin se hace especialmente difcil. Desde la comunidad cientfica se
arrojan una gran cantidad de datos, muchos de ellos no demasiado coincidentes entre si.
Estudios recientes indican por ejemplo que de las aproximadamente 200.000 toneladas de
mercurio emitidas a la atmsfera desde 1890, aproximadamente el 95% se encuentran en
suelos terrestres, alrededor de un 3% en aguas ocenicas y un 2% en la atmsfera. Tambin
se estima que entre el 40 y el 75% del mercurio atmosfrico actual tiene como origen
fuentes antropognicas.
5. W.F. Fitzgerald, R.P. Mason and G.M. Vandal, Water, Air and Soil Pollution (1991), 56, 745-767
6. J.O. Nriagu, The Biogeochemistry of Mercury in the Environment, Elsevier, New York, (1979)
11
Introduccin
Partiendo de estos valores, Mason et al.7 proponen los siguientes diagramas de flujo
post y pre era Industrial (Figura 1.1 a y b respectivamente).
a
3
Flujos en 10 Kg/ao
3
Reservas en 10 Kg
b
3
Flujos en 10 Kg/ao
3
Reservas en 10 Kg
Figura 1.1. Diagrama de flujo de mercurio post (a) y pre (b) era industrial
7. R.P. Mason, W.F. Fitzgerald and F.M.M. Morel, Geochimica et Cosmochimica Acta (1994), 58, 3191-3198
12
Introduccin
Ec. 1.1
Ec. 1.2
Ec. 1.3
Ec. 1.4
Aunque el Hg(II) producido por el ozono pueda ser devuelto a la forma Hg(0) a partir de
la reaccin con sulfito, la segunda de las reacciones (Ec. 1.2) tiene una cintica mucho
ms rpida por lo que el resultado final es la acumulacin de Hg(II). Este tipo de
mecanismos abren una va a la posible deposicin del mercurio elemental presente en la
atmsfera.
13
Introduccin
14
Introduccin
15
Introduccin
mercurio elemental que satura las capas superiores y acaba finalmente volatilizndose a la
atmsfera.
8. J.J. Rytuba, Mercury Geoenvironmental Models, in Progress on geoenvironmental models for selected mineral deposit
types, R.R. Seal and N.K. Foley (ed.), (2002), [en lnea], <http://pubs.usgs.gov/of/2002/of02-195/>, [5 Marzo 2004]
9. EPA Mercury Study Report to Congress, Volume II, An Inventory of Anthropogenic Mercury Emissions in the United States,
(1997), [en lnea], <http://www.epa.gov/ttn/oarpg/t3/reports/volume2.pdf>, [5 Marzo 2004]
16
Introduccin
Industrias cloro-lcali con ctodo de mercurio. Pese a estar siendo sustituidas por
otras tecnologas, existen todava en todo el mundo un gran nmero de este tipo de
industrias. En ellas se produce NaOH y Cl2 a partir de una salmuera de elevada
pureza. El costo energtico es muy elevado, existiendo adems riesgos graves para
el medio ambiente por la utilizacin de mercurio. Como contrapartida, los
productos obtenidos a partir de este mtodo son de una elevada pureza y
concentracin. En el apartado 1.11.1 se explicarn ms detalles acerca de este
proceso.
Pilas. Aunque son muchos los pases que han prohibido su utilizacin, existe todava
un uso bastante generalizado de pilas o bateras a partir de celdas de Zn/Hg. En
ellas, el ctodo es HgO y el nodo Zn, siendo la reaccin global la mostrada en la
Ec. 1.5:
Zn + HgO ZnO + Hg
Ec. 1.5
C 2H
HgCl
2
+ HCl
CH
HgSO
= CHCl
4
+ H 2 O
CH 3CHO
Ec. 1.6
Ec. 1.7
17
Introduccin
Uno de los usos ms extendidos del mercurio es el de relleno dental. Una aleacin
de Ag(66.7-74.5%), Sn(25.3-27.0%), Cu(0.0-6.0%) y Zn(0.0-1.9%) es mezclada con
mercurio elemental (aproximadamente 1:1 en peso) para dar un amalgama dental.
La matriz final es dura, resistente a la abrasin, se adhiere fuertemente al diente,
tiene una muy baja solubilidad y es impermeable a la saliva.
10. G. Lfroth, Methylmercury, a review hazards and side effects associated with the emission of mercury compounds into
natural systems, Stockholm, (1970)
18
Introduccin
1.5.1.1
Fuentes de combustin
Calderas. Las calderas son utilizadas en un sinfn de mbitos, tanto para obtener
electricidad como para obtener calor. As, se encuentran en instalaciones pblicas y
privadas, en industrias y tambin a escala mucho ms pequea (pero con un uso
muy extendido) en los hogares de muchas familias. Pueden ser alimentadas con
carbn, petrleo, gas natural o combinacin de ellos.
Incineradoras
de
residuos
peligrosos.
Son
utilizadas
bsicamente
para
el
1.5.1.2
Fuentes de produccin
19
Introduccin
parmetros.
stos
incluyen
termmetros,
sensores
de
presin
1.5.1.3
Fuentes diversas
20
Introduccin
Tabla 1.1. Listado de fuentes de rea en los Estados Unidos para el periodo 1994-1995
Fuentes
Rotura de lmparas
1.0
0.7
Preparaciones dentales
0.4
Disposicin en vertederos
0.1
Fuentes mviles
Uso de pinturas
11. Environmental Health Criteria 101, Methylmercury, World Health Organisation, Geneva (1990)
21
Introduccin
confeccionar el pan, el cual haba sido tratado con un funguicida basado en este
compuesto organometlico de mercurio.
Toda esta serie de desastres sensibilizaron a la comunidad internacional, de tal forma
que la Organizacin Mundial de la Salud (OMS) public en 1977 el primer libro de la serie
Environmental
Health
Criteria
(Critres
dhygine
de
lenvironnement)
dedicado
12
22
Introduccin
Introduccin
24
Introduccin
25
Introduccin
26
Ciertamente,
existen
otras
especies
de
mercurio
(la
mayora
tambin
Introduccin
17. J.R.D. Guimaraes, M. Roulet, M. Lucotte and D. Mergler, The Science of the Total Environment (2000), 261, 91-98
18. N. Mikac et al., Applied Organometallic Chemistry (1999), 13, 715-725
19. E.J. Zillioux, D.B. Porcella and J.M. Benoit, Environmental Toxicology and Chemistry (1993), 12, 2245-2264
20. Sung-Chan Choi, T. Chase Jr. and R. Bartha, Applied and Environmental Microbiology (1994), 60(11), 4072
21. P. Povari and M. Verta, Water Air and Soil Pollution (1995), 80, 765-773
22. C.C. Gilmour, E.A. Henry and R. Mitchell, Environmental Science and Technology (1992), 26(11), 2281-2287
27
Introduccin
23. J.K. King et al., Environmental Science and Technology (2001), 35(12), 2491-2496
24. J.M. Benoit, C.C. Gilmour and R.P. Mason, Environmental Science and Technology (2001), 35(1), 127-132
25. R. Ebinghaus and R.D. Wilken, Applied Organometallic Chemistry (1993), 7, 127-135
26. R. Ebinghaus, H. Hintelmann and R.D. Wilken, Fresenius J. Analytical Chemistry (1994), 350, 21-29
28
Introduccin
+ 2H 2 O + 4H
3CH
+ CO
+ 4Hg
2+
+ 4H
Ec. 1.8
2
4
+ CH 3 Hg
+ 3H
H 2 S + CO
+ Hg
2+
+ 2H 2O
Ec. 1.9
27. R.S. Oremland, C.W. Culbertson and M.R. Winfrey, Applied and Environmental Microbiology (1991), 57(1), 130-137
28. M.C. Marvin-DiPasquale and R.S. Oremland, Environmental Science and Technology (1998), 32, 2556-2563
29. B. Michalke, Fresenius Journal of Analytical Chemistry (1999), 363, 439-445
30. R.Lobinski, Spectrochimica Acta B (1998), 53, 177-185
29
Introduccin
Definir las especies objeto del anlisis desde un punto de vista qumico. La
asociacin de un elemento con la fraccin correspondiente a una etapa del proceso
u operacin de laboratorio (p.e. plomo extrable en tolueno) no debe ser
denominado anlisis de especiacin.
Definir el marco adecuado para el grupo de especies que van a ser analizadas. No
tiene sentido hablar de la determinacin de especies de un elemento cuando se
est intentando analizar nicamente una de esas especies. Este es sin embargo un
error muy comn en el campo de la especiacin de mercurio, donde un gran
nmero de metodologas tratan de determinar nicamente el metilmercurio.
30
Introduccin
queremos analizar. El mayor problema suele venir dado por la estabilidad cintica, que se
encuentra afectada por equilibrios cido-base, redox y de complejacin. En estos casos,
pequeos cambios en las condiciones ambientales pueden modificar los equilibrios fsicoqumicos existentes de manera que las especies finalmente determinadas no representan a
aquellas que se encontraban originalmente en la muestra. El conocimiento de este hecho
es crucial, aunque a menudo no ayuda al analista a solucionar el problema, pues en
muchas ocasiones no se puede hacer literalmente nada para evitarlo.
32. G.E. Brown Jr., A.L. Foster and J.D. Ostergren, Proc. Natl. Acad. Sci. (1999), 96, 3388-3395
33. S.Z. Lee, L. Chang, C.M. Chen, Y.I. Tsai and M.C. Liu, Water Science and Technology (2001), 43(2), 187-196
34. U. Skyllberg, K. Xia, P.R. Bloom, E.A. Nater and W.F. Bleam, Journal of Environmental Quality (2000), 29, 855-865
35. Y. Yin, H.E. Allen, C.P. Huang and P.F. Sanders, Soil Science (1997), 162(1), 35-45
36. D. Wallschlger, M.V.M. Desai, M. Spengler, C.C. Windmller and R.D. Wilken, Journal of Environmental Quality (1998),
27, 1044-1054
37. Y. Yin, H.E. Allen, C.P. Huang and P.F. Sanders, Analytica Chimica Acta (1997), 341, 73-82
38. A.Amirbahman, A.L. Reid, T.A. Haines, J.S. Kahl and C. Arnold, Environmental Science and Technology (2002), 36, 690695
31
Introduccin
que depende, entre otras variables, del pH) o la posibilidad de formacin de fases
coloidales39 (0.4m-10kDa).
Existen diferentes enfoques a la hora de trabajar con el concepto de especiacin de
mercurio en suelos y sedimentos. Cada uno de ellos afronta el problema desde un punto de
vista distinto, llegndose tambin a resultados finales que permiten grados de
conocimiento del sistema muy diferentes. Estos enfoques pueden resumirse en31:
-
utilizan
estas
tcnicas
persigue
la
identificacin
cuantificacin
de
39. C.L. Babiarz, J.P. Hurley, S.R. Hoffman, A.W. Andren, M.M. Shafer and D.E. Armstrong, Environmental Science and
Technology (2001), 35, 4773-4782
40. M.O. Barnett et al., Environmental Science and Technology (1997), 31(11), 3037-3043
32
Introduccin
Especiacin directa. Los mtodos directos apareceran como una buena opcin
para la determinacin de especies de un mismo elemento. El hecho de prevenir la
extraccin de estas especies elimina el riesgo de modificacin de las mismas
durante el proceso. Existen sin embargo escasas tcnicas directas que puedan ser
aplicadas a slidos, y an menos, a matrices tan complejas como suelos y
sedimentos. Para el caso de aguas, se han aplicado con xito tanto tcnicas
electroqumicas como las clsicas UV-vis o fluorimetra, aunque stas tambin
fallan en matrices complejas (p.e. fluidos biolgicos) debido a su falta de
sensibilidad y selectividad. En el caso de suelos, se han aplicado algunos mtodos
fsicos (sobretodo relacionados con radiacin X) obtenindose buenos resultados, si
bien tal y como se ver en la seccin 1.8.1.2 presentan tambin inconvenientes
importantes.
1.8.1.1
41. A. Ivask, K. Hakkila and M. Virta, Analytical Chemistry (2001), 73, 5168-5171
42. A. Ivask, Proceedings of the Senspol Technical Meeting on Sensors for Characterisation and Monitoring of Contaminated
Sites (2002) Sevilla
43. M.Hempel, Y.K. Chau, B.J. Dutka, R. McInnis, K.K. Kwan and D. Liu, Analyst (1995), 120, 721-724
33
Introduccin
1.8.1.1.1
Extraccin
44. S. Kitamura, T. Tsukamura, K. Hayakawa, K. Sumino and T. Shibata, Medical Biology (1966), 72, 274-280 (en japons)
45. K. Sumino, Kobe Journal of Medical Science (1968), 14, 115-130
46. G. West, Acta Chemica Scandinavica (1968), 22, 2277-2280
47. M. Hempel, H. Hintelmann and R.D. Wilken, Analyst (1992), 117, 669-672
48. A.M. Carro, E. Rub, M.H. Bollain, R.A. Lorenzo and R. Cela, Applied Organometallic Chemistry (1994), 8, 665-676
49. R.A. Lorenzo, M.J. Vzquez, A.M. Carro and R. Cela, Trends in Analytical Chemistry (1999), 18(6), 410-416
34
Introduccin
50. H. Hintelmann and R.D. Wilken, The Science of the Total Environment (1995), 166, 1-10
51. R.D. Wilken and H. Hintelmann, Water, Air, and Soil Pollution (1991), 56, 427-437
52. H. Hintelmann, R. Falter, G. Ilgen and R.D. Evans, Fresenius Journal of Analytical Chemistry (1997), 58, 363-370
53. A. Alli, R. Jaff and R. Jones, Journal of High Resolution Chromatography (1994), 17, 745-748
54. M. Filippelli, Chemosphere (1999), 39(7), 1199-1210
55. J. Qian et al., Fresenius Journal of Analytical Chemistry (2000), 367, 467-473
56. L. Lambertsson, E. Lundberg, M. Nilsson and W. Frech, Journal of Analytical Atomic Spectrometry (2001), 16, 1296-1301
57. C.R. Hammerschmidt and W.F. Fitzgerald, Analytical Chemistry (2001), 73, 5930-5936
58. C.J. Cappon and J.C. Smith, Analytical Chemistry (1977), 49(3), 365-369
59. C.J. Cappon and J.C. Smith, Bulletin of Environmental Contamination and Toxicology (1978), 19, 600-607
60. C.J. Cappon and J.C. Smith, Analytical Chemistry (1980), 52, 1527-1529
61. R. Falter, H. Hintelmann and P. Quevauviller, Chemosphere (1999), 39(7), 1039-1049
62. M.J. Vzquez, A.M. Carro, R.A. Lorenzo and R. Cela, Analytical Chemistry (1997), 69, 221-225
63. C.M. Tseng, A. De Diego, F.M. Martin and O.F.X. Donard, Journal of Analytical Atomic Spectrometry (1997), 12, 629-635
35
Introduccin
objeto
de
algunas
crticas
por
la
posible
formacin
artificial
de
36
Introduccin
Estos datos pueden ser tomados como referencia a la hora de valorar las tendencias
actuales en la etapa de extraccin de las especies de mercurio dentro del denominado pretratamiento de muestra.
71. P. Lansens, C. Meuleman, M. Leermakers and W. Baeyens, Analytica Chimica Acta (1990), 234, 417-424
72. P. Quevauviller, M. Filippelli and M. Horvat, Trends in Analytical Chemistry (2000), 19(2+3), 157-166
37
Introduccin
1.8.1.1.2
Separacin y determinacin
selectiva73),
la
mayora
de
tcnicas
separativas
utilizadas
son
cromatogrficas.
La cromatografa de gases (GC) es la tcnica cromatogrfica ms usada de cuantas se
aplican a especiacin de mercurio desde que West propusiera su mtodo a finales de los
aos 60. Inicialmente se usaban columnas de longitud y dimetro interno variable (15-30 m
y 0.3-0.75 cm respectivamente) empacadas con OV 101 adsorbido sobre un soporte slido
de Chromosorb W-HP74,75 o columnas de vidrio silanizado empacadas con DEGS-PS al 5% 76.
Este tipo de columnas presenta sin embargo una serie de problemas53:
-
El carcter polar del enlace Hg-haluro puede interaccionar fuertemente con la columna,
dando lugar a picos cromatogrficos con importantes colas, incluso a aparentes reacciones
de descomposicin77,78. Por ello, se hace necesaria la pasivacin de la columna, que se
lleva a cabo mediante acondicionamientos con soluciones de cloruro de mercurio(II),
yoduro de mercurio(II) o yoduro de metoxietilmercurio79. Esta pasivacin no tiene adems
un carcter permanente, sino que debe irse realizando peridicamente.
Con la intencin de superar estos inconvenientes, se han evaluado diferentes tipos de
columnas capilares durante la ltima dcada utilizando tanto fases polares como fases no
polares. Las columnas con mayor grosor de fase estacionaria y baja polaridad han
demostrado ser las ms adecuadas. Este tipo de columnas ofrece mayor eficiencia en las
separaciones y mejores resoluciones, an y siendo necesaria la inyeccin de volmenes de
muestra pequeos (alrededor de 1l).
Una caracterstica de la GC es la necesidad de trabajar con analitos voltiles. En
muchas ocasiones se procede a la inyeccin directa de los haluros organomercuriales
73. M.C.R. Rezende, R.C. Campos and A.J. Curtius, Journal of Analytical Atomic Spectrometry (1993), 8, 247
74. C.J. Cappon and J.C. Smith, Bulletin of Environmental Contamination and Toxicology (1978), 19(5), 600-607
75. S. Rapsomanikis and P.J. Craig, Analytica Chimica Acta (1991), 248(2), 563-567
76. J.E. Snchez Ura and A. Sanz-Medel, Talanta (1998), 47, 509-524
77. J.E. OReilly, Journal of Chromatography, (1982), 238, 433
78. E. Bulska, D.C. Baxter and W. Frech, Analytica Chimica Acta (1991), 249, 545-554
79. A.M. Carro and M.C. Mejuto, Journal of Chromatography A (2000), 882, 283-307
38
Introduccin
Formacin del alquilderivado tras la extraccin de los analitos de la matriz (se han
ensayado tanto la etilacin como la butilacin). Algunos estudios mencionan sin
embargo la posible formacin artificial de metilmercurio a causa de las impurezas
presentes en el tetraetilborato61 empleado en la alquilacin.
Otro aspecto muy importante es el sistema de deteccin usado. Durante mucho tiempo,
el acoplamiento cromatgrafo de gases detector de captura electrnica (GC-ECD) ha sido
el ms utilizado48,62,81, y es todava incluido en la configuracin experimental de algunos
trabajos82. Debido a la baja selectividad del detector ECD, es necesario el empleo de
etapas previas de limpieza (como la propuesta por West). En la mayora de casos, la
inyeccin de la muestra se hace en fase tolueno o benceno.
Uno de los detectores especficos ms populares es el detector de emisin atmica con
plasma inducido por microondas (MIP-AES). La deteccin usando emisiones atmicas es
ventajosa en el caso de los metilelementos, debido a su intensa emisin y baja seal de
fondo que proporciona una excelente sensibilidad y un alto grado de selectividad. En la
mayora de casos en los que se usa el sistema GC-MIP-AES, la muestra es inyectada previa
derivatizacin de los analitos (alquilacin)55,66,83, por lo que los problemas habidos en la
separacin de los haluros de mercurio, la necesidad de pasivar la columna, etc. quedan
superados. En algn caso, el sistema GC-MIP-AES se acopla tambin a la introduccin de
muestra por espacio de cabeza71 (HeadSpace, HS), procedindose entonces a la adicin de
cido yodoactico y cido sulfrico para facilitar la volatilizacin de los analitos. La
mezcla se termostatiza a 80C durante 5 minutos para luego ser introducida en el sistema
HS-GC-MIP-AES con una inyeccin de entre 15 y 30 segundos.
El acoplamiento cromatgrafo de gases plasma por acoplamiento inducido
espectrmetro de masas (GC-ICP-MS) va ganando adeptos poco a poco y apunta como una
de las tcnicas de futuro en la especiacin de mercurio por su gran sensibilidad y
selectividad. En la mayora de casos se da una derivatizacin previa de los analitos para
80. W. Liu and H.K. Lee, Journal of Chromatography A (1999), 834, 45-63
81. R. Cela et al., Environmental Technology (1992), 13, 11-22
82. R. Liu et al., Environmental Pollution (2003), 124, 39-46
83. J. Snell, J. Qian, M. Johansson, K. Smit and W. Frech, Analyst (1998), 123, 905-909
39
Introduccin
84. H. Hintelmann and R.D. Evans, Fresenius Journal of Analytical Chemistry (1997), 358, 378-385
85. M. Horvat, L. Liang and N.S. Bloom, Analytica Chimica Acta (1993), 282, 153-168
86. A. De Diego, C.M. Tseng, T. Stoichev, D. Amouroux and O.F.X. Donard, Journal of Analytical Atomic Spectrometry (1998),
623-629
87. C.F. Harrington, Trends in Analytical Chemistry (2000), 19(2+3), 167-179
88. D.M. Snchez, R. Martn, R. Morante, J. Marn and M.L. Munuera, Talanta (2000), 52, 671-679
89. S. Ro-Segade and C. Bendicho, Talanta (1999), 48, 477-484
90. H. Hintelmann, M. Hempel and R.D. Wilken, Environmental Science and Technology (1995), 29, 1845-1850
40
Introduccin
41
Introduccin
1.8.1.1.3
Ya hemos comentado que las tcnicas hbridas son las tcnicas analticas de rutina
utilizadas para la determinacin de especies de mercurio en matrices medioambientales
(suelos y sedimentos entre otros). Precisamente por ello, los qumicos analticos que
trabajan en este campo se han preocupado y mucho del control de calidad de los anlisis
que se realizan. El punto clave para la especiacin exitosa es precisamente mantener
informacin acerca de las especies a analizar en todas las etapas del proceso. Las fuentes
de error ms comunes son29:
-
Deteccin final: todava hoy, los errores de calibracin siguen jugando un papel
clave en la obtencin de resultados errneos. Igualmente, errores procedentes de
interferencias espectrales y de masa, superposicin de picos o mala correccin del
fondo siguen siendo habituales.
100. T.H. Lee and S.J. Jiang, Analytica Chimica Acta (2000), 413, 197-205
101. M. Silva, A.B. Soldado, E. Blanco-Gonzlez and A. Sanz-Medel, Journal of Analytical Atomic Spectrometry (2000), 15,
513-518
42
Introduccin
Laboratorio: una educacin adecuada del personal que realiza los anlisis es bsica
en la obtencin de buenos resultados. No slo es necesaria la bsqueda de buena
reproducibilidad, sino que precisin y correccin deben ser igualmente exigidas.
NIES, en Japn.
NRCC, en Canad.
IAEA, en Austria.
La Tabla 1.272 muestra los CRMs preparados hasta el momento por estas instituciones
en el campo de la especiacin de mercurio:
102. B. de Guillebon, F. Pannier, F. Seby, D. Bennink and P. Quevauviller, Trends in Analytical Chemistry (2001), 20(3), 160166
43
Introduccin
Productor
SRM 1974
NIST
SRM 2974
NIST
SRM 2976
NIST
CRM 463
103,104
BCR
CRM 464
103,104
BCR
BCR
DORM-1
NRCC
LUTS-1
NRCC
NIES
IAEA-350
IAEA
IAEA-356
IAEA
IAEA-140
IAEA
IAEA-142
IAEA
IAEA-085
IAEA
IAEA-086
IAEA
1.8.1.2
CRM 580105-107
NIES No.13
44
Introduccin
se suelen situar en torno a valores porcentuales del peso total de la muestra. Esto hace
que
sean
descartados
como
fuente
de
informacin
en
muchas
aplicaciones
Radiacin X
Objetivo de anlisis
Identificacin de estructuras cristalinas
presentes en la muestra. Determinacin
por comparacin con compuestos de
referencia
Haz de e-
45
Introduccin
Aunque los lmites de deteccin se han reducido mucho respecto a las tcnicas
convencionales basadas en tubos de rayos-X, en muchas ocasiones siguen siendo
demasiado altos.
1.8.1.3
Desorcin trmica
mediante
el
acoplamiento
de
una
unidad
calefactora
controlada
46
Introduccin
Hg0
Hg2Cl2
170
HgCl2
HgO
270-535, 160-495
HgS
Hg en Pirita
450
Hg en Esfarelita
600
Hg-Substancias hmicas
200-300
Como puede observarse en la Tabla 1.4, los intervalos dados para algunos compuestos
son bastante grandes, e incluso para alguno de ellos se determinan diferentes
temperaturas de desorcin. La causa de esto es precisamente uno de los grandes
inconvenientes de la tcnica, y es que la temperatura de desorcin depende de una serie
de variables no siempre controlables, como la matriz, el tamao de partcula o la
presencia de irregularidades en la red cristalina. Esto hace de la termodesorcin una
tcnica cualitativa-semicuantitativa, que puede llegar a ser muy til si es complementada
adecuadamente con otras tcnicas o metodologas36,122.
1.8.1.4
Extraccin secuencial
47
Introduccin
del reactivo o disolvente usado para su extraccin (formas extractables en EDTA, en cido
cetico 0.1M, etc.). As, en este sentido, la extraccin secuencial aporta informacin sobre
la prediccin de comportamiento de una muestra en diferentes medios, hecho que genera
informacin del tipo de unin entre las especies metlicas y el sustrato del suelo.
Uno de los grandes problemas de los procedimientos de extraccin secuencial es la falta
de reproducibilidad de los resultados. Por un lado, existe una gran variedad de
procedimientos de extraccin secuencial (incluso mtodos especficos para el estudio de
determinados elementos, entre ellos el mercurio123-126). Probablemente los procedimientos
de Tessier127 y BCR-SES128,129 sean, sin embargo, los ms utilizados en la actualidad,
constando de cinco y tres etapas de extraccin secuencial respectivamente (ver Tabla
1.5):
Tabla 1.5. Procedimiento de extraccin secuencial definido por Tessier y BCR-SES
BCR-SES
Tessier
Fraccin
Condiciones de extraccin
3ml 0.02M HNO3 + 2mL 30% H2O2 (pH 2), 85C, 2h; 3mL
30% H2O2 (pH 2), 85C, 2h; 5mL 3.2M CH3COONH4 in
20% HNO3 + 7mL H2O, 25C, 30 min
123. P.J. Lechler, J.R. Miller, L.C. Hsu and M.O. Desilets, Journal of Geochemical Exploration (1997), 58, 259-267
124. D. Wallschlger, M.V.M. Desai, M. Spengler and R.D. Wilken, Journal of Environmental Quality (1998), 27, 1034-1044
125. M.M. Incio, V. Pereira and M.S. Pinto, Geoderma (1998), 85, 325-339
126. N.S. Bloom et al., Analytica Chimica Acta (2003), 479(2), 233-248
127. A. Tessier, P.G.C. Campbell and M. Bisson, Analytical Chemistry (1979), 51, 844-851
128. P. Quevauviller et al., The Science of the Total Environment (1997), 205, 223-234
129. P. Quevauviller et al., Report EUR 17554 EN European Comission (1997) Brussels
48
Introduccin
49
Introduccin
tres los pilares interdependientes y sinrgicos del desarrollo sostenible. As, en la divisin
para el Desarrollo Sostenido de Naciones Unidas, se incluyen ms de treinta categoras,
como parte de dichos pilares132: Agricultura, Atmsfera, Biodiversidad, Biotecnologa,
Capacidad de construccin, Patrones de consumo y produccin, Demografa, Ocanos y
mares, Sequa y desertizacin, Educacin y conocimiento, Energa, Finanzas, Bosques,
Aguas terrestres, Salud, Asentamientos humanos, Indicadores, Industria, Informacin para
la toma de decisiones y participacin, Toma de decisiones integrada, Cooperacin
internacional
para
la
capacitacin
medioambiental,
Ley
internacional,
Acuerdos
50
Introduccin
Las categoras generales de impacto ambiental que suelen (deben) ser consideradas
incluyen el uso de recursos, la salud humana, as como las consecuencias sobre el medio
ambiente. De este modo, LCA puede ser de gran utilidad en:
-
Interpretacin
Debe ser sin embargo remarcado que LCA se encuentra todava en una etapa temprana
de desarrollo. Algunas fases de LCA como la Valoracin del impacto ambiental deben
recorrer an un largo camino para su completo perfeccionamiento. As pues, es siempre
importante que los resultados obtenidos por LCA sean interpretados y aplicados
apropiadamente de acuerdo a sus limitaciones. De entre stas, destacaremos:
-
51
Introduccin
Por otro lado, LCA es una de tantas herramientas de gestin medioambiental (entre
ellas
gestin
de
riesgo,
evaluacin
del
desarrollo
medioambiental,
auditorias
52
Introduccin
Cambiadores inorgnicos.
134. K. Dorfner, Ion Exchangers, Walter der Gruyter Publisher, Berln, (1991)
135. F. Helfferich, Ion Exchange, Dover publications, New York, (1995)
53
Introduccin
1.10.1
(entre
otras
opciones)
que
conferirn
diferentes
grados
de
de
los
intercambiadores
inicos.
Considerando
este
grupo,
los
54
Introduccin
1.10.2
Caractersticas fsicas. La densidad de las perlas de resina vara entre 1.1 y 1.4
g/ml, en funcin del grado de hidratacin de la resina. El tamao de grano y su
homogeneidad son factores importantes a controlar, pues condicionarn las
separaciones y la uniformidad del empaquetamiento de la columna.
55
Introduccin
1.10.3
+o
R B + A
+o
Ec. 1.10
en la que pasado un cierto tiempo, que oscila entre unos minutos y unos das, se alcanza el
equilibrio. La constante que rige el equilibrio se denomina coeficiente de selectividad, y es
representada por la letra E con dos subndices: uno inferior, referido a la especie por la
que el intercambiador experimenta una menor afinidad, y otro superior, que se reserva
para la especie por la que ste es ms afn (Ec. 1.11):
E
B
A
C Aq B
CBqA
Ec. 1.11
siendo C la concentracin de los iones en fase acuosa y q en fase resina. C y q pueden ser
expresados en distintas unidades, resultando en consecuencia valores tambin distintos
para el coeficiente de selectividad. Tratndose de la constante de equilibrio de la
reaccin, queda claro que cualquier coeficiente estequiomtrico que aparezca en la
reaccin, se ver igualmente reflejado en la expresin del coeficiente de selectividad
como exponente.
136. M. Valcrcel, A. Gmez Hens, Tcnicas Analticas de Separacin, Revert, Barcelona, (1994)
56
Introduccin
X AY B
X A (1 Y A )
=
YB X A
Y A (1 X A )
Ec. 1.12
donde X corresponde a la fraccin molar de los iones en fase acuosa y Y a la fraccin molar
de los iones en fase resina. En el caso del factor de separacin, a diferencia del
coeficiente de selectividad, no aparecen los exponentes correspondientes a coeficientes
estequiomtricos (en caso de existir). As, y E solo coincidirn cuando trabajemos con
reacciones de estequiometra 1 a 1.
al
intercambio
inico
en
un
137. D. Muraviev, R. Khamizov and N.A. Tikhovov, Solvent Extraction and Ion Exchange (1998), 16(1), 151-221
57
Introduccin
intercambio inico convencional. Esta otra reaccin ser la que realmente desplace el
conjunto del proceso hacia una regeneracin eficiente.
1.10.4
Si aceptamos esta simplificacin del proceso, el intercambio inico nos aparece como
un fenmeno de trasferencia de masa en lugar de como una reaccin qumica. Son as
diferenciables las siguientes etapas:
-
58
Introduccin
Sin embargo, el flujo de los dos iones debe estar acoplado: la electroneutralidad debe
ser conservada, y esto requiere que la transferencia de carga originada por el
desplazamiento de uno de los iones deba ser compensada con una transferencia de carga
equivalente del otro in en sentido contrario. Adems, para la mayora de casos, la
reaccin de intercambio inico es lo suficientemente rpida como para no ser la
determinante de la velocidad. Esto nos llevar a considerar tan slo dos etapas cinticas:
-
En muchas ocasiones una cintica controlada por transferencia de masa en fase resina,
ser denominada como un control por difusin intraparticular. Esta terminologa es
correcta en este caso, dado los impedimentos que la matriz pone a que la conveccin
ocurra. De este modo, el principal causante del movimiento de los iones dentro de la fase
resina ser la difusin y por tanto la cintica vendr controlada por este factor.
Ciertos errores deben ser admitidos en la aproximacin de asimilar el intercambio inico
como una simple transferencia de masa. En ciertas ocasiones por ejemplo, el intercambio
puede ir acompaado de reacciones qumicas138. Ejemplos tpicos son la neutralizacin de
un intercambiador catinico en forma protnica por una base, o la complejacin de un
metal de transicin por una resina quelatante139,140. Evidentemente, estas reacciones, de
darse, pueden afectar a la cintica del intercambio.
1.10.5
138. M. Nativ, S. Goldstein and G. Schmuckler, Journal of Inorganic and Nuclear Chemistry (1975), 37, 1951-1956
139. A.I. Kalinichev et al., Journal of Inorganic and Nuclear Chemistry (1981), 43, 787-789
140. J.L. Cortina et al., Reactive and Functional Polymers (1998), 38, 269-278
141. V. Gorshkov, D. Muraviev and A. Warshawsky, Solvent Extraction and Ion Exchange (1998), 16(1), 1-73
59
Introduccin
60
Introduccin
elevada pureza.
61
Introduccin
1.11.1
Hidrxido de sodio144,145
143. D. Muraviev, R.K. Khamizov and N.A. Tikhonov in Advances in Ion Exchange for Industry and Reserach, P.A. Williams and
A. Dyer (ed.), Royal Society of Chemistry, Cambridge, (1999)
144. R.E. Kirk and D.F. Othmer, Encyclopedia of Chemical Technology, M. Howe-Grant (ed.), Jonhn Wiley & Sons, New York,
(1991)
145. Ullmanns Encyklopdie der technischen Chemie, Verlag Chemie, Weinheim-Deerfield BeachBessel, (1966)
62
Introduccin
A lo largo de todos estos aos, el cloro ha sido la especie que ha marcado el ritmo de
produccin de la industria cloro-lcali, puesto que de los productos finales que se
obtienen, es el que posee un mayor valor aadido. As pues, el aporte al mercado
internacional de NaOH ha venido muy influenciado por la demanda que este mercado haca
de cloro. sta ha sido la principal causa de las grandes fluctuaciones en el precio del NaOH
a lo largo de toda su historia146. En la actualidad, an y encontrndose en claro declive, el
cloro sigue siendo el producto considerado como ms importante de esta industria, por lo
que el NaOH sigue sometido a las fluctuaciones en su demanda.
En cuanto a las aplicaciones del hidrxido sdico, stas son muchas y variadas, pero
derivan principalmente de su habilidad como base fuerte para reaccionar con un gran
nmero de sustancias, incluyendo aluminio y zinc. Adems, muchos de los compuestos de
sodio formados a partir de sosa son solubles en agua. As, algunos de los usos ms
importantes del NaOH147 son: neutralizacin de cidos, industria del aluminio, produccin
de rayn y celulosa, industria del papel, industria textil, produccin de petrleo y
refinados, detergentes, explosivos, industria farmacutica, etc.
La aplicacin final de una determinada sosa depender tanto de su concentracin como
de su pureza. Estos factores vendrn fijados por el mtodo de produccin utilizado. As por
ejemplo, es posible encontrar soluciones de sosa al 50%, al 50% purificada, al 50% grado
membrana o al 73%, entre otros, siendo cada una de ellas apropiada para diferentes
procesos148.
146. K. Gilbert, Industrial Inorganic Chemicals: Production and Uses, R. Thompson (ed.), The Royal Society of Chemistry,
London, (1995)
147. PPG company, [en lnea], <http://www.ppg.com/chm_Chloralk/chlor_home_prod.htm>, [5 Marzo 2004]
148. Dow chemicals company [en lnea], <http://www.dow.com>, [5 Marzo 2004]
149. Publicacin de Solvay, Complejo Solvay Martorell, (1997)
63
Introduccin
Una vez purificada, la salmuera es decantada primero y filtrada despus para eliminar
de este modo cualquier resto de slido presente en la disolucin.
Otro reactivo a tener en cuenta es la energa elctrica150. La energa elctrica es
realmente una materia prima de la electrlisis, ya que una gran parte de esta energa se
encuentra dentro de los productos finales, cuyo nivel energtico es muy superior al de los
reactivos de partida.
c. 21% membrana
b
Figura 1.4. Cuotas de produccin en la sntesis industrial de NaOH
Es importante remarcar que estos valores son a nivel mundial, y que esta distribucin
vara mucho en funcin del pas considerado. As, en Japn, y a consecuencia del desastre
de Minamata, las celdas de mercurio fueron yendo substituidas primero por las celdas de
diafragma, y a partir de 1980 por celdas de membrana de manera que a da de hoy,
aproximadamente un 90% de la produccin proviene de celdas de membrana. Ninguna de
las industrias cloro-lcali en ese pas funciona con celda de mercurio152. En la Unin
Europea siguen dominando las industrias con celda de mercurio (55% a Junio de 2000),
150. Solvay company [en lnea], <http://www.solvay.com>, [5 Marzo 2004]
151. M.G. Beal, Modern Chlor-Alkali Technology (Volume 6), R.W. Curry (ed.), The Royal Society of Chemistry, Cambridge,
(1995)
152. J. Thornton, Pandoras Poison: Chlorine, Health, and a New Environmental Strategy, MIT Press, Massachusetts, (2000)
64
Introduccin
1.11.1.2.1
Como ya ha sido comentado, todos los procesos actuales de produccin de sosa son
electrolticos. Esto implica la presencia de un nodo donde se oxidan unas especies y un
ctodo donde se reducen otras. El proceso de ctodo de mercurio fue introducido a finales
del siglo XIX por Castner-Kellner (1892), y el hecho que lo diferencia de otros es como su
nombre indica el ctodo, que es de mercurio. Este ctodo ser el responsable de que la
reduccin que aqu se produzca sea la del Na+/Na y no la del H+/H2. Ello es debido
sobretodo a los sobrepotenciales que se dan a la reduccin de H+/H2, y al hecho que el
sodio metlico formado da lugar a una amalgama muy estable con el mercurio lquido,
actuando as como fuerza desplazante de la reaccin. La segunda etapa del proceso es la
reaccin de la amalgama con agua para dar hidrgeno y sosa. Las reacciones implicadas en
este proceso pueden observarse en la Ec. 1.13.
nodo
Ctodo
: Cl
Reaccin
Balance
0 . 5Cl
: xHg + Na
en Ctodo
+e
+e
NaHg
: NaHg
+ H 2 O 0 . 5H
+ NaOH
+ 0 . 5H
+ xHg
Ec. 1.13
+ NaOH
153. European Comission [en lnea], Integrated Pollution Prevention and Control (IPPC); Reference Document on Best
Available Techniques in the Chlor-Alkali Manufacturing Industry (December 2001), <www.envir.ee/ippc/docs/chloralkali.doc>, [5 Marzo 2004]
154. P.J. Paine, Compliance with chlor-alkali mercury regulations, 1986-1989 status report, Minister of Supply and Services
Canada, (1994)
65
Introduccin
1.11.1.2.2
Proceso de diafragma
0 . 5H
+ OH
Ec. 1.14
Permite la circulacin a travs suyo de la solucin acuosa (y por lo tanto los iones
Na+ y Cl-), pero no las pequeas burbujas de gas, evitando as la mezcla del
hidrgeno y el cloro. Sin embargo, no puede evitar el paso del cloro disuelto
(alrededor de un 4%) del nodo al ctodo.
1.11.1.2.3
155. R. Stringer and P. Johnston, Chlorine and the Environment. An overview of the chlorine industry, Kluwer Academic
Publishers, London (2001)
156. F.Y. Masuda, Journal of Applied Electrochemistry (1986), 16, 317-331
157. H.R. Bolton, Journal of Chemical Technology and Biotechnology (1992), 54, 341-347
66
Introduccin
separa las cmaras del ctodo y nodo. En este caso sin embargo, la barrera es
impermeable al agua y conductora de iones.
Las reacciones implicadas son tambin las mismas que en el caso del proceso de
diafragma. En ausencia de corriente, el agua no es transportada de una a otra cmara,
pero una vez aplicamos corriente, las pequeas cantidades de agua que se encuentran
solvatando a los iones son transportadas. La membrana permite tan slo el paso de los
cationes (Na+) de la cmara andica a la catdica, y es por esto que para este tipo de
proceso es requerida una salmuera inicial de elevada pureza (incluso ms pura que para el
caso del ctodo de mercurio), pues pequeas impurezas de calcio pueden inducir la
precipitacin de ste en la superficie de la membrana en forma de Ca(OH)2. El desarrollo
de una membrana estable en condiciones electrolticas, con altas concentraciones salinas,
elevado gradiente de pH entre cmaras y presencia de agentes oxidantes fuertes ha sido el
principal problema en este tipo de procesos.
Las ventajas del mtodo son:
-
1.11.2
67
Introduccin
+ Ca (OH ) 2 CaCO
+ 2LiOH
Ec. 1.15
68
Introduccin
1.12 OBJETIVOS
En la determinacin de especies de mercurio:
-
69
Introduccin
Comparacin de procesos.
70
Metodologa y Experimental
1A
71
Metodologa y Experimental
2 Metodologa y Experimental
1B
2C
72
Metodologa y Experimental
cido
10 mL HNO3 concentrado
Temperatura
Presin
tamizadas
(tomndose
la
fraccin
inferior
100
m)
homogeneizadas.
-
Predigestin de 10 h.
73
Metodologa y Experimental
2.2 ESPECTROSCOPIA
DE
EMISIN
ATMICA
EN
PLASMA
POR
2. A. Woller, H. Garraud, F. Martin, O.F.X. Donard and P. Fodor, Journal of Analytical Atomic Spectrometry (1997), 12, 53-56
3. J. Allibone, E. Fatemian and P.J. Walker, Journal of Analytical Atomic Spectrometry (1999), 14, 235-239
4. E. Fatemian, J. Allibone and P.J. Walker, The Analyst, 124, 1233-1236
5. D.A. Skoog and J.J. Leary, Anlisis Instrumental, McGraw-Hill, (1994)
6. A. Varma, CRC Handbook of Inductively Coupled Plasma Atomic Emission Spectroscopy, CRC Press, Boca Ratn, (1990)
7. G.L. Moore, Introduction to Inductively Coupled Plasma Atomic Emission Spectroscopy, Analytical Spectroscopy Library
(Vol.3), Elsevier, (1989)
8. P.W.J.M. Boumans, Inductively Coupled Plasma Emission Spectroscopy. Part I: Methodology, Instrumentation and
Performance, John Wiley & Sons, New York, (1987)
74
Metodologa y Experimental
Na
Ca
Mg
Hg
670.78
589.59
766.49
422.67
279.806
180.731
194.227
0.0018
0.069
0.060
0.010
0.015
0.02
0.025
V, Ti
Te, Ti, V
Ti
Fe
Al, V
Interferencias
Al igual que para la digestin, las muestras fueron secadas a temperatura ambiente
hasta peso constante, molturadas, tamizadas (tomndose la fraccin inferior a 100 m) y
homogeneizadas.
En la bibliografa se describe la posible degradacin del metilmercurio por la accin de
la luz11,12. Por ello, y aunque los tiempos de extraccin son cortos, se trabaj siempre con
tubos de extraccin protegidos con papel de aluminio.
9. R.K. Winge, V.A. Fassel, V.J. Peterson and M.A. Floyd, Inductively Coupled Plasma-Atomic Emission Spectroscopy. An Atlas
of Spectral Information, Elsevier, Amsterdam, (1985)
10. R.A. Lorenzo, M.J. Vzquez, A.M. Carro and R. Cela, Trends in Analytical Chemistry (1999), 18(6), 410-416
11. X. Gaona and M. Valiente, Analytica Chimica Acta (2003), 480, 219-230
75
Metodologa y Experimental
2.4 ELECTROFORESIS
CAPILAR
DE
ALTA
RESOLUCIN
(HIGH
--
toma
elctrica
N
N
++
Flujo electrosmtico
tampn
NODO (+)
tampn
CTODO (-)
12. I. Suda, M. Suda and K. Hirayama, Archives of Toxicology (1993), 67, 365-368
13. D.R. Baker, Capillary Electrophoresis, John Wiley & Sons, New York, (1995)
76
Metodologa y Experimental
Puesto que todos los analitos son transportados a travs del capilar, aniones y
cationes pueden ser separados en un solo anlisis.
77
Metodologa y Experimental
Moles
Molaridada
Absorbancia UV/vis
10-13-10-16
10-5-10-7
Absorbancia indirecta
10-12-10-15
10-4-10-6
Fluorescencia
10-15-10-17
10-7-10-9
Fluorescencia indirecta
10-14-10-16
10-6-10-8
10-18-10-20
10-13-10-16
Espectrometra de masas
10-16-10-17
10-8-10-10
Amperometra
10-18-10-19
10-7-10-10
Conductividad
10-15-10-16
10-7-10-9
ndice de refraccin
10-14-10-16
10-6-10-8
Radiomtrico
10-17-10-19
10-10-10-12
78
Metodologa y Experimental
79
Metodologa y Experimental
2.4.3.1
Stacking, FASS)
Esta es la tcnica ms sencilla de preconcentracin en lnea. Los efectos de inyectar la
muestra en matrices de baja conductividad fueron explicados primeramente por Mikkers et
al17. En general, este mtodo se basa en la idea que los iones que migran
electroforticamente desde una solucin de baja conductividad a otra de alta
14. J.P. Quirino and S. Terabe, Journal of Chromatography A (2000), 902, 119-135
15. Z.K. Shihabi, Journal of Chromatography A (2000), 902, 107-117
16. D.M. Osbourn, D.J. Weiss and C.E. Lunte, Electrophoresis (2000), 21, 2768-2779
17. F.E.P. Mikkers, F.M. Everaerts and T.P.E.M. Verheggen, Journal of Chromatography (1979), 169, 11-20
80
Metodologa y Experimental
2.4.3.2
18. R.L. Chien and D.S. Burgi, Analytical Chemistry (1992), 64, 489A-496A
19. R.L. Chien and D.S. Burgi, Analytical Chemistry (1992), 64, 1046-1050
81
Metodologa y Experimental
2.4.3.3
Un
nmero
de
pequeos
pptidos
naturales
sintticos
pueden
ser
2.4.4 Aplicaciones
Electroforesis capilar es aplicable a la determinacin de una gran variedad de
compuestos que van desde simples iones inorgnicos e iones metlicos, hasta complejas
biomolculas incluyendo oligosacridos, cidos nucleicos y protenas.
En teora, cualquier serie de compuestos que difiera en su relacin carga/radio puede
ser separada por Electroforesis Capilar de Zona. Compuestos que varen en su grado de
particin entre una miscela y un tampn pueden ser separados por Cromatografa Capilar
Miscelar Electrocintica. Si las molculas tienen la misma relacin carga/radio, pero
tienen tamaos diferentes, pueden ser separados por Electroforesis Capilar de Gel,
mientras que si los compuestos difieren en su pI, pueden ser separados por Focalizacin
Isoelctrica Capilar.
Separaciones quirales pueden ser llevadas tambin a cabo en electroforesis capilar
normalmente mediante la utilizacin de Ciclodextrinas (CD). Los analitos forman complejos
de inclusin con las CD, con la parte hidrofbica enclavada en el interior de la CD,
mientras que la parte polar del analitos interacciona con los grupos polares hidroxil de los
que tambin dispone la CD. La separacin del racmico vendr dada por la diferencia en
las constantes de formacin de los complejos entre la CD y los enantimeros D y L.
82
Metodologa y Experimental
pH del tampn.
Potencial aplicado.
Potencial aplicado
Tiempo de anlisis
Volumen de inyeccin
Deteccin
Temperatura del capilar
Soluciones de lavado
83
Metodologa y Experimental
25. J.I. Goldstein et al., Scanning Electron Microscopy and X-Ray Microanalysis, 2nd edition, Plenum Press, New York, (1992)
84
Metodologa y Experimental
LA
ESPECIACIN
DE
MERCURIO
EN
SUELOS
CONTAMINADOS
Como ya se apuntaba en el apartado 1.8.1.2 de la Introduccin, las tcnicas de
absorcin de rayos-X que utilizan como fuente de radiacin instalaciones sincrotrn pueden
aportar informacin muy valiosa en el campo de la especiacin en matrices
medioambientales. Con sus ventajas e inconvenientes, lo cierto es que tcnicas como
EXAFS o XANES se encuentran entre las tcnicas ms potentes en el reconocimiento de
diferentes especies de un mismo elemento gracias a su especificidad de seal del entorno
atmico. En los siguientes apartados se introducir la teora tanto del fenmeno fsico de
la radiacin sincrotrn como de las propias tcnicas XAFS, y se describirn y discutirn
algunos de los detalles del procedimiento experimental.
P =
2 e 2 2 dp
3 m 02c 3 dt
Ec. 2.1
26. Synchrotron Radiation: Earth, Environmental and Material Sciences Applications, G. Henderson and D.R. Baker (ed.),
Mineralogical Association of Canada, Otawa, (2002)
27. X-Ray Absorption: Principles, Applications, Techniques of EXAFS, SEXAFS and XANES, D.C. Koningsberger and R. Prins
(ed.) John Wiley & Sons, New York, (1988)
28. European Synchrotron Radiation Facility (ESRF) [en lnea], <http://www.esrf.fr>, [5 Marzo 2004]
29. A. Thompson et al., X-Ray Data Booklet, University of California Berkeley, California, (2001)
85
Metodologa y Experimental
E
=
m 0c 2
1
1
v
c
Ec. 2.2
Ec. 2.3
sin
( + cos )
Ec. 2.4
El ngulo de abertura (la mitad del ngulo de luz sincrotrn divergente por encima y
debajo del plano de rbita) es 1/. 1/ es una expresin muy importante pues
determina la divergencia del haz de fotones. As, a mayor energa de los electrones, menor
ngulo de apertura de la radiacin sincrotrn generada, menor divergencia y por lo tanto
mayor luminosidad.
86
Metodologa y Experimental
Radio de curvatura
Radio de curvatura
Fuerza de
aceleracin
Fuerza de
aceleracin
Trayectoria
del electrn
Electrn
Electrn
Trayectoria del electrn
Campo de radiacin
Campo de radiacin
a.
b.
Ec. 2.5
Figura 2.3. rbita del electrn donde se muestra el diferencial de momento correspondiente a
la variacin de ngulo d
Para electrones relativistas, v c, podemos sustituir dp/dt por la Ec. 2.1, y sustituir
por E/moc2 y pc por E. Se obtiene entonces la energa irradiada en cada vuelta del electrn
cuando es desviado por un imn dipolar con un radio de curvatura (ver Ec. 2.6).
P =
2 e 2c
E4
3 (m 0c 2 ) 4 2
Ec. 2.6
87
Metodologa y Experimental
Si el radio es el mismo a lo largo de toda la rbita, la perdida de energa (E) por electrn
y por vuelta en forma de radiacin sincrotrn puede ser calculada segn:
E =
Pdt
orbita
=P
Ec. 2.7
Sustituyendo Ec. 2.6 en Ec. 2.7, y usando las unidades adecuadas, se obtiene la expresin:
E [keV ] = 88.5
E 4 [GeV
[m ]
Ec. 2.8
La Ec. 2.8 muestra como la energa perdida es la cuarta potencia de la energa del
electrn. La radiacin sincrotrn empieza a ser apreciable para electrones con energa de
unas pocas decenas de MeV y significativa para GeV. As, una mayor radiacin sincrotrn es
producida en anillos de alta energa.
La radiacin sincrotrn es una fuente de luz extremadamente verstil, til en un gran
nmero de investigaciones experimentales. Esa versatilidad le es conferida por una serie
de interesantes propiedades, entre las cuales destacan:
-
88
Metodologa y Experimental
89
Metodologa y Experimental
Segunda generacin:
Finales 70 principios 90
Tercera generacin:
Principios 90 - presente
Cuarta generacin:
En desarrollo
90
Desarrollo relevante
Emisin de luz por electrones no relativistas (frecuencia de Larmor).
Descubrimiento de los rayos-X por Rntgen.
Principio del sincrotrn (McMillan 1945, Veksler 1945).
Primera observacin de radiacin sincrotrn (GE Research Laboratory en
Schnectady, Nueva York, Abril 1947).
Primeros experimentalistas usando aceleradores de electrones para llevar a cabo
experimentos con radiacin sincrotrn. La RS sigue siendo una molestia para los
fsicos de altas energas.
Inicio del uso regular de la RS en investigacin (DESY en Alemania, Daresbury en
Reino Unido y NBS en los E.U.A.). A finales de los 60, se construyen pequeos
sincrotrones. A principios de los 70, experimentos parsito empiezan y
rpidamente extienden en instalaciones de alta energa (SSRL en Stanford).
Nacimiento de una slida comunidad de usuarios.
La investigacin con RS se extiende por todo el mundo. Instalaciones VUV y
anillos de rayos-X son fundados y construidos para ser dedicados exclusivamente
a la RS. nfasis en la fuente de imanes (dipolares) de curvatura. La comunidad
de usuarios se expande a un gran nmero de disciplinas.
Insercin de aparatos con alternancia peridica de campos magnticos en las
secciones lineales. Construccin durante los noventa de anillos basados en este
tipo de inserciones (SRRC, APS, ALS, ESRF, Spring8). Siguen implementndose en
la actualidad.
Aplicaciones especiales: p.e. Lser de electrones libres, aceleradores lineales de
recuperacin de energa y otros.
Metodologa y Experimental
2.6.2.1
Dispersin (scattering)
herramientas
para
el
estudio
2.6.2.2
Absorcin
fotoelectrnica
(PhotoElectron
Spectroscopy,
PES)
fenmenos
91
Metodologa y Experimental
Electrones de energa definida pueden ser usados como bistur para fragmentar
molculas con especificidad espacial. Procesos qumicos foto-inducidos pueden ser ideados
para aplicaciones litogrficas.
Otra aplicacin de los procesos de absorcin es el uso de un micro-haz de radiacin-X
ajustable en energa para microscopia o incluso terapia por radiacin. El ajuste de la
energa del electrn por debajo y por encima de la energa de un determinado mximo de
absorcin puede dar lugar a imgenes de contraste de una alta resolucin muy apropiada
para aplicaciones mdicas. Incluso, con el desarrollo de los submicro-haces, la radiacin
sincrotrn puede llegar a ser una herramienta muy til en la caracterizacin estructural de
nanomateriales.
= ln
I0
It
Ec. 2.9
donde Io e It son las intensidades del haz de radiacin-X incidente y del haz transmitido a
travs de la muestra respectivamente. Esta frmula coincide con las respectivas relaciones
para coeficientes de absorcin lineal en otros rangos espectrales (UV, visible, IR), ya que la
absorcin en el rango de los rayos-X tambin sigue la ley de Bougher-Lambert-Beer.
El coeficiente de absorcin de rayos-X aumenta bruscamente para determinados valores
de energa de radiacin incidente, efecto que se superpone con un suave pero constante
decrecimiento de la curva (E) con el aumento de la energa. Esta absorcin de resonancia
inducida por la fotoionizacin de tomos de ciertos elementos define formas triangulares
para las bandas de absorcin. En las proximidades del mximo de absorcin, (E)
manifiesta una estructura fina, esto son ligeras oscilaciones que se distinguen con claridad
del fondo de absorcin. Estas oscilaciones se dan porque cuando el fotn es absorbido (ver
Figura 2.5), un fotoelectrn es excitado desde un nivel electrnico cercano al ncleo del
tomo absorbente (crculo negro) hasta un estado prcticamente de libertad (mostradas
como circunferencias alrededor del crculo negro). Cuando la energa del fotn excitado
30. Y.V. Zubavichus and Y.L. Slovokhotov, Russian Chemical Reviews (2001), 70, 373-403
31. X-Ray Absorption Fine Structure, S.S. Hasnain (ed.), Ellis Horwood, Chichester, (1991)
92
Metodologa y Experimental
k =
2m (E E 0 )
Ec. 2.10
h2
donde E y E0 son la energa del fotn y la energa de enlace que la onda del fotoelectrn
experimenta, respectivamente.
tomo vecino
h
Electrn dispersado
Fotoelectrn
Figura 2.5. Representacin esquemtica del origen del fenmeno XAFS
La onda del fotoelectrn es entonces dispersada por los electrones de un tomo vecino
(crculo gris), propagndose partes de dicha onda de nuevo hacia el centro del tomo
absorbente. Como el estado inicial origen del fotoelectrn se encuentra muy localizado en
las cercanas del centro del tomo absorbente, slo las componentes que sean dispersadas
haca ese punto van a influir en la probabilidad de absorcin. Una vez llegadas a ese punto,
las ondas dispersas pueden encontrarse en fase o desfasadas respecto a la onda generada
por el tomo absorbente, y por lo tanto magnificar o inhibir la probabilidad de absorcin.
Esto es, la interferencia entre el fotoelectrn saliente y la onda dispersa de regreso es el
origen de las oscilaciones en el espectro.
La estructura fina se subdivide en dos tipos (ver Figura 2.6): absorcin de rayos-X por
estructuras en las proximidades del salto electrnico (X-ray Absorption Near Edge
Structure, XANES), y absorcin de rayos-X de estructura fina extendida (Extended X-ray
Absorption Fine Structure, EXAFS). XANES cubre un rango de energa desde 50eV antes del
mximo de absorcin hasta unos 100-150eV despus de dicho mximo. Esta estructura fina
esta compuesta de bandas de resonancia relativamente estrechas generadas por
transiciones electrnicas desde niveles atmicos internos hasta niveles de energa
93
Metodologa y Experimental
Absorbancia normalizada
100
EXAFS
80
60
40
XANES
20
0
12.0
12.2
12.4
Energia (eV)
12.6
12.8x10
Figura 2.6. Espectro de las zonas XANES y EXAFS para el HgO. Anlisis en modo transmitancia.
Espectro corregido en energa, desviaciones pre y post-Edge y normalizado
94
Metodologa y Experimental
If
I0
Ec. 2.11
32. Introduction to EXAFS Experiments and Theory, curso impartido por Bruce Ravel, Washington, (2000)
95
Metodologa y Experimental
Io
Muestra
IT
IR
ptica
Fuente
IF
Amplificadores y
Ref.
Barrera
de Pb
Ordenador
Figura 2.7. Representacin esquemtica de una lnea de haz dedicada a experimentacin XAFS
Slit 1
Monocromador
Segundo espejo
Slit 3
Primer espejo
Slit 2
96
Metodologa y Experimental
97
Metodologa y Experimental
4.5 GeV
Dos disponibles: Si(1,1,1) y Si(3,1,1)
0.5-1 eV
2mm x 3mm
Detector/es de transmitancia
Detector de fluorescencia
ngulo haz-muestra
Temperatura de trabajo
6 GeV
Dos disponibles: Si(1,1,1) y Si(2,2,0)
0.5-1 eV
>1013 fot./seg.
200m x 80m
Detector/es de transmitancia
Detector/es de fluorescencia
ngulo haz-muestra
Temperatura de trabajo
45
Ambiente
98
Metodologa y Experimental
99
Metodologa y Experimental
Regeneracin de la resina
100
Metodologa y Experimental
101
Metodologa y Experimental
33. koinventare fr Verpackungen Band I, Schriftenreihe Umwelt Nr. 250/I, Bundesamt fr Umwelt, Wald und Landschaft
(BUWAL), (1994)
102
Resultados y Discusin
1A
2B
103
Resultados y Discusin
104
Resultados y Discusin
Este captulo recoge aquellos aspectos ms relevantes de los resultados de cada uno de
los estudios presentados. Sin embargo, no se pretende repetir el contenido de los trabajos
que se adjuntan en los anexos, sino que se intenta dar una visin de conjunto resaltando
aquellos aspectos que suponen una nueva contribucin en el contexto del mercurio como
contaminante global.
105
Resultados y Discusin
Resultados y Discusin
y Ec. 3.2) y la presencia de cisteina (ver Ec. 3.3 y Ec. 3.4). En este segundo caso, el
verdadero causante de la degradacin sera un producto de degradacin de la cisteina, la
cistina. Este efecto aparece descrito en la bibliografa para otros derivados de la cisteina1.
O 2(sol ) + h 1O 2
Ec. 3.1
O 2 + HgCH 3+ / HgC 6 H 5+ Hg 2 +
Ec. 3.2
N H2
N H2
2x
+O2
SH
HOOC
HOOC
L-cisteina
Ec. 3.3
N H2
L-cistina
L-cystine
L-cysteine
COOH
R
S
N H2
H2N
N H2
S
HOOC
Hg2+
COOH
SH
HOOC
HgR+
Ec. 3.4
HS
R
COOH
N H2
el
desarrollar
una
tcnica
de
preconcentracin
que
permitiese
la
107
Resultados y Discusin
Puesto que el nmero de variables a considerar para desarrollar el mtodo era elevado,
se procedi en primer lugar a realizar una seleccin de aquellas variables que realmente
eran significativas para el resultado final. Para ello, se realiz un diseo de experimentos
binario reducido en tres rdenes, de manera que el nmero de experimentos a realizar fue
27-3. Con este tipo de reduccin, los experimentos necesarios pasan de ser 128 a 16,
teniendo tan slo como contrapartida la confusin de factores principales con
interacciones de tercer orden.
Las variables escogidas para este primer diseo de seleccin se mostraron ya en el
apartado 2.4.5 del captulo Experimental, si bien la Tabla 3.1 muestra adems los dominios
para cada una de ellas.
Tabla 3.1. Variables y dominios considerados para el diseo de experimentos 27-3IV realizado
Variable
Relacin NaCl/Acetonitrilo
Dominio
0.6 1
0.1 1 ppm
0.02 0.2 mM
% rellenado de capilar
8 20 %
pH tampn
8.2 8.6
Potencial aplicado
8 14 kV
Alcohol utilizado
Metanol Propanol
108
Resultados y Discusin
Valor / Dominio
0.7 ppm
pH tampn
8.6
Alcohol utilizado
Propanol
Potencial aplicado
10 25 kV
0.001 0.01 mM
15 20 %
% Acetonitrilo en muestra
30 50 %
0.2 0.4 %
El diseo utilizado para la optimizacin de estas cinco variables fue tipo cubo centrado
en las caras, que con un nmero reducido de experimentos permite la incorporacin de
tres niveles distintos para cada una de las variables consideradas. As, el nmero final de
experimentos fue de 26, ms tres puntos centrales. El efecto de cada una de las variables,
tigation: 021120_5V ar (M LR)
as como de las interacciones Inves
secundarias,
puede observarse en la Figura 3.1.
S c aled & Centered Coeffic ients for HgM e A rea
2000
1000
-1000
%C F * Sal
C TB* Sal
C TB* % C F
Pot* Sal
Pot* % C F
Pot* C TB
C TB* C TB
Pot* Pot
Sal
Ac t
%C F
C TB
Pot
-2000
109
Resultados y Discusin
Una vez encontrado el modelo, la maximizacin del rea de metilmercurio arroja como
valores ptimos de las variables:
-
% acetonitrilo: 30.1%
Kcounts
y = 0,6753x + 0,1034
2
R = 0,9992
y = 0,2353x + 0,2413
2
R = 0,9975
y = 0,1983x + 0,2161
2
R = 0,9999
10
C Hg (ppm)
Figura 3.2. Curvas de calibrado para HgCH3Cl (!), HgCl2 (!) y Hg(NO3)2 (") en medio acuoso y
15 ppm Au
110
Resultados y Discusin
y = 0,2638x + 0,2777
R2 = 0,9997
5
Kcounts
y = 0,2367x + 0,132
R2 = 0,999
y = 0,3294x + 0,1541
R2 = 0,9953
y = 0,7223x + 0,2584
R2 = 0,9963
1
0
0
C(ppm)
Figura 3.3. Curvas de calibrado para HgCH3Cl: en medio cisteina (!), en medio Au (") y
Hg(NO3)2: en medio cisteina ("), en medio Au (!)
De este modo, el siguiente sistema de ecuaciones (ver Ec. 3.5 - Ec. 3.10) nos permite la
determinacin directa de metilmercurio y mercurio inorgnico de una muestra
desconocida, a partir de su emisin en medio Au y en medio cisteina, y a partir de las
cuatro curvas de calibrado individuales de HgCH3Cl y Hg(NO3)2 en medio cisteina y Au.
AHgMe / Au = aC HgMe + b
Ec. 3.5
Ec. 3.6
AHg / Au = fC Hg + g
Ec. 3.7
AHg /Cist = hC Hg + j
Ec. 3.8
A Au = AHgMe / Au + AHg / Au
Ec. 3.9
Ec. 3.10
111
Resultados y Discusin
0.5 ppm).
En relacin con el segundo de los inconvenientes, se est trabajando en el acoplamiento
de columnas de preconcentracin al sistema descrito, para permitir la determinacin de
mercurio a niveles mucho ms bajos.
3.2 DETERMINACIN
DE
ESPECIES
DE
MERCURIO
EN
SUELOS
112
Resultados y Discusin
113
Resultados y Discusin
114
Resultados y Discusin
ambas muestras son 170 y 560 ppm respectivamente. Cuando se intent obtener el
espectro XAFS de las muestras, si bien ambas eran muy ruidosas por la baja concentracin
de mercurio presente, no se pudo obtener ninguna estructura fina (ver Figura 3.4) para la
100
Fluorescencia Normalizada
80
60
40
20
0
-20
12.0
12.2
12.4
E nergy (eV )
12.6
12.8x 10
Energa (eV)
Figura 3.4. Espectro de Fluorescencia de la muestra TP1. Gris: 10 espectros individuales. Negro:
media de los espectros individuales
La explicacin a este hecho se encuentra en los datos mostrados en la Tabla 3.3, donde
aparecen reflejadas las concentraciones de As, Zn, Pb, Cu y Fe presentes en ambas
muestras.
Tabla 3.3. Concentracin de As, Zn, Pb, Cu y Fe en las muestras MIN1 y TP1
Elemento
As
169 4
4600 900
Zn
151 14
4300 100
Pb
23 2
22500 400
Cu
11 1
2100 100
Fe
28400 500
43100 700
115
Resultados y Discusin
6 000
Cuentas
4 000
As
Zn
5 000
Fe
3 000
2 000
Cu
Zn
Hg
1 000
0
3 40
4 40
5 40
640
740
C an al
Figura 3.5. Ventana de fluorescencia completa para la muestra TP1. Energa de la radiacin-X
incidente: 12300eV
Tal y como se puede observar en la Figura 3.5, Zn y As interfieren con la seal de Hg por
solapamiento de alguna de sus lneas de fluorescencia con la lnea de fluorescencia de
mercurio analizada. Por otro lado, la presencia ms que habitual de Fe en muestras de
suelo, puede llegar a afectar tambin a la calidad de la seal, al dar lugar a una seal de
fondo grande, si se encuentra en concentraciones tambin grandes. Para detectores de
fluorescencia muy sensibles, puede llegar a producir incluso la saturacin del detector.
Lnea de haz A1, HASYLAB, Hamburgo. Proyecto I-01-061EC, Abril 2002, 6 das de
anlisis. Optimizacin de las condiciones de haz para anlisis de mercurio en
laboratorios sincrotrn. Estudio de sistemas binarios y terciarios de mercurio.
Lnea de haz A1, HASYLAB, Hamburgo. Proyecto I-01-061EC, Agosto 2002, 7 das de
anlisis. Estudio de sistemas binarios y terciarios de mercurio. Anlisis de muestras
de suelo reales.
116
Resultados y Discusin
Lnea de haz ID26, ESRF, Grenoble. Proyecto ME-599, Marzo de 2003, 6 das de
anlisis. Determinacin de especies de mercurio en muestras ultradiluidas.
Posibilidades de una instalacin sincrotrn de 3 generacin para el anlisis de
muestras medioambientales. Caracterizacin de suelos de mina. Estudios de
absorcin de HgMeCl y HgMeOH sobre diferentes componentes de suelo.
Lnea de haz A1, HASYLAB, Hamburgo. Proyecto I-02-099 EC, Diciembre de 2003, 6
das de anlisis. Comportamiento y distribucin espacial de las especies de
mercurio en muestras mineras.
Lnea de haz A1, HASYLAB, Hamburgo. Proyecto I-03-043EC, Abril de 2004, 12 das
de anlisis. Comportamiento y distribucin espacial de las especies de mercurio en
muestras mineras
117
Resultados y Discusin
H2O
Energa solar
CaO
H2O
(40C)
(H<0)
Suspensin de
Ca(OH)2(s) a 60C
Ca(OH)2(s) Ca2+ + 2OH-
(Kps)
Separacin de fases
Ca(OH)2(s)(exc)
MOH (producto)
R2Ca
MCl + M2SO4
(tecnologa IXISS)
Subproducto
Figura 3.6. Esquema del proceso diseado para la produccin de hidrxidos alcalinos basado en
el uso de resinas de intercambio inico
118
Resultados y Discusin
119
Resultados y Discusin
Na), el aumento en la fuerza inica da lugar a una perdida en la afinidad de la resina por
Ca. El efecto del pH no fue significativo en el intervalo de acidez estudiado.
0,3
CNaOH(M)
0,25
0,2
0,15
0,1
0,05
293K
313K
333K
353K
0
0
10
20
30
40
50
60
70
80
90
t(min)
Figura 3.7. Cinticas de la reaccin de sntesis de NaOH en reactor discontinuo y a distintas
temperaturas de trabajo
Estos datos sirvieron para tener un mayor conocimiento del sistema (donde pH i I varan
a lo largo del proceso) y para fijar un valor de temperatura ptimo teniendo en cuenta
consideraciones cinticas y termodinmicas. Como se ha comentado, mayor temperatura
implica mayor selectividad de la resina respecto a calcio (y por lo tanto mayor
concentracin final de MOH). Sin embargo, a partir de 333K la velocidad de reaccin
disminuye tal y como se observa en la Figura 3.7. Adems, se debe considerar el coste
econmico de calentar grandes cantidades de disolucin en el reactor. Por todo ello, se
fijo 333K como valor de compromiso para la temperatura de trabajo.
El aumento de la variable m/V supuso un aumento en la concentracin final de producto
(NaOH), puesto que la cantidad de in Na+ presente en el sistema era mayor (ver Figura
3.8). Sin embargo, en trminos de grado de conversin de resina, ste experiment un
decrecimiento con el aumento de la variable m/V (ver Figura 3.9). Este hecho es lgico y
se correlaciona perfectamente con lo observado para la concentracin, pues al aumentar
la fraccin molar de Na en disolucin, la afinidad de la resina por el Ca aumenta,
restringiendo su liberacin al medio.
120
Resultados y Discusin
0.8
0.7
CMOH(M)
0.6
0.5
0.4
0.3
0.2
LiOH
NaOH
KOH
1.5
0.1
0
0
0.5
2.5
mq/V
Figura 3.8. Variacin de la concentracin final de producto en funcin de la cantidad de resina
en la mezcla de reaccin para los sistemas Li+/Ca2+, Na+/Ca2+ y K+/Ca2+. T=293K. q: capacidad de
la resina para cada uno de los cationes alcalinos considerados
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
LiOH
NaOH
KOH
0
0
0.5
1.5
2.5
mq/V
Figura 3.9. Grado de conversin de la resina en funcin de mq/V para los tres metales de
estudio. q: capacidad de la resina para cada uno de los cationes alcalinos considerados
121
Resultados y Discusin
122
Resultados y Discusin
'
Ca
'
M
) C M ) (K ps
qo
C M4 + ((K ps
m
Ca
M
)=0
V
Ec. 3.11
123
Resultados y Discusin
Es importante mencionar que en todo momento los datos son referidos a la produccin
de 1000 kg de NaOH puro, o en su defecto 2000 kg de NaOH al 50%.
Energa
para
la
energa
puesta en marcha,
Transporte
Total [MJ]
Cantidad
[MJ]
Cantidad
[MJ]
[MJ]
Electricidad
9590
1200 kWh
4320
6 kWh
20
13930
Carbn
160
38 kg
1150
0 kg
1310
Derivados del
petrleo
180
27 kg
1160
7 kg
280
1620
Gas natural
190
92 m3
3680
0 m3
10
3880
Total
10120
310
20740
10310
Tabla 3.5. Materia prima utilizada para la produccin de la energa necesaria en la sntesis de
1000 kg de NaOH
Lignito
160 kg
120 m3
Hulla
230 kg
77 kg
Uranio (mineral)
13 g
710 m3
(energa hidroelctrica)
124
Mineral de hierro
0.46 kg
Piedra caliza
10.5 kg
Sal
590 kg
Arena
0.2 kg
Agua de proceso
5.3 m3
Resultados y Discusin
Suministro de
energa
Combustin +
proceso
Transporte
TOTAL
500
2580
20
3100
Benceno [g]
1.2
1.2
PAH [g]
0.017
0.017
5.9
5.9
0.018
0.018
0.0004
0.0004
Metano [g]
2400
2400
650
3380
70
4100
CO2 [g]
100000
1000000
20000
1120000
CO [g]
50
630
20
700
NH3 [g]
HF [g]
N2O [g]
5.5
5.5
HCl [g]
85
65
1000
8900
100
10000
500
6500
200
7200
150
125
Resultados y Discusin
que podran ser considerados a la hora de disear un proceso global ms respetuoso con el
medio ambiente.
El clculo energtico del proceso fue el que tuvo un error asociado ms grande. As, en
este inventario no se consideraron detalles de ingeniera como tipo de reactores, energas
de agitacin, energas de iniciacin, etc. sino que se tomaron nicamente aquellos valores
de energas obtenidos directamente de los clculos termodinmicos. Esta aproximacin
debe ser tenida en cuenta a la hora de extraer las conclusiones finales.
Se definieron en primer lugar los reactivos, ya que la produccin de stos implica a su
vez la utilizacin de energa y generacin de residuos, que deben ser igualmente tenidos
en cuenta en el balance global. De este modo, se consideraron nicamente dos reactivos:
CaO y Na2SO4. Se aproxim vida ilimitada de la resina, siendo plenamente conscientes que
esta aseveracin es incorrecta (el promedio de tiempo de vida de estas resinas comerciales
suele ser de 5 aos si se utilizan de forma continuada). BUWAL 1994 nos aporta los datos
en cuanto a la obtencin de estos reactivos. As, para la obtencin de CaO se referencia el
siguiente balance de materia (Figura 3.10):
Requerimientos energticos:
- 1096 MJ energa trmica (100% gas natural)
- 6.9 kWh energa elctrica
388 kg CaCO3
194 kg CaO
Calcinacin en horno
Figura 3.10. Balance de materia para la produccin de CaO a partir de la calcinacin de CaCO3
El reactivo de partida en este caso fue CaCO3, que es obtenido por extraccin directa,
de manera que el inventariado para el CaO finaliz en este punto.
El segundo de los reactivos utilizados en el proceso (en este caso para la regeneracin
de la resina) es el Na2SO4. De nuevo consultando con BUWAL 1994, se obtuvo el diagrama
de flujos de la Figura 3.11. En este caso, tanto NaCl como H2SO4 proceden de sendos
procesos industriales, y por tanto fue tambin necesaria la obtencin de su balance
ecolgico. Para el caso del NaCl, ste se dispuso directamente a partir de BUWAL 1994 (ver
126
Resultados y Discusin
Tabla 3.8 a Tabla 3.11), mientras que para el caso del cido sulfrico es referenciado un
nuevo balance de materia (ver Figura 3.12).
Requerimientos energticos:
- 211.2 kg fuel oil
- 156 kWh energa elctrica
1974 kg NaCl
2400 kg Na2SO4
3670 kg HCl (31%)
1725 kg H2SO4
Proceso de NaCl
Figura 3.11. Balance de materia para la sntesis de Na2SO4 a partir de NaCl y H2SO4
Energa puesta en
Transporte
Total [MJ]
energa
marcha, [MJ]
Cantidad
[MJ]
Cantidad
[MJ]
Electricidad
430
58 kWh
210
640
Carbn
30
4 kg
130
160
Derivados del
petrleo
70
11 kg
470
0.9 kg
40
580
Gas natural
90
39 m3
1580
1670
Total
620
40
3050
2390
Tabla 3.9. Materia prima utilizada para la produccin de la energa necesaria en la sntesis de
1000 kg de NaCl
Lignito
7.3 kg
43 m3
Hulla
16 kg
15 kg
Uranio (mineral)
0.47 g
45 m3
127
Resultados y Discusin
14 kg
Sal
1070 kg
Arena
0.06 kg
Agua de proceso
4.6 m3
Suministro de
energa
Combustin +
proceso
Transporte
TOTAL
40
280
Benceno [g]
0.18
0.18
PAH [g]
0.0012
0.0012
0.51
0.51
0.0036
0.0036
0.00002
0.00002
Metano [g]
370
370
130
1500
CO2 [g]
17000
155000
3000
175000
CO [g]
83
90
NH3 [g]
0.074
0.074
HF [g]
0.45
0.45
N2O [g]
0.44
0.44
HCl [g]
4.2
5.8
100
980
20
1100
100
1370
30
1500
320
1630
10
Requerimientos energticos:
0 MJ
570 kg S
Agua
Catalizador
Combustin, seguido de
1725 kg H2SO4
proceso de contacto
128
Resultados y Discusin
Azufre y agua aparecen como los nicos reactivos de partida para la sntesis de cido
sulfrico. El azufre es obtenido por extraccin directa (no por ningn proceso industrial),
por lo que el inventariado de ambos procesos finaliz aqu.
Es necesario remarcar que, para aquellos balances de materia en los que los trminos
energticos no aparecen desglosados, esta etapa debe hacerse manualmente para una
mayor exactitud en el resultado final. Este desglose energtico viene marcado por el pas
en el que nos encontremos, puesto que cada pas tiene su propia distribucin de fuentes
energticas. La Tabla 3.12 muestra los desgloses energticos para una serie de pases
europeos. Para realizar todos los clculos incluidos en este estudio, se utilizaron los
porcentajes correspondientes a Espaa.
Hulla
(%)
Lignito
Petrleo
(%)
Gas N.
(%)
Energa at.
(%)
Hidroelctrica
(%)
Otros
(%)
Blgica
22.8
2.2
14.2
58.4
2.4
Alemania
28.9
18.8
1.2
4.5
39.3
6.3
Dinamarca
73.8
0.1
3.7
2.1
5.2
12.7
2.4
Francia
7.3
0.4
2.1
1.5
72.9
15.7
0.1
Gran Bretaa
59
8.2
3.2
26.4
2.9
0.3
Italia
8.6
0.5
43.4
14.5
8.2
23.5
1.3
Luxemburgo
25.7
14.7
1.7
10.3
37.2
9.7
0.7
Holanda
30.4
3.1
3.6
48.9
12.8
0.9
0.3
Portugal
36.3
0.7
43.4
0.4
2.6
16.6
Espaa
30.5
9.9
9.2
1.7
35.4
13.3
Suiza
4.7
1.4
2.5
1.2
50
40.1
0.1
Austria
7.1
8.8
2.8
9.9
6.3
65
0.1
Noruega
0.1
0.2
0.2
0.3
99.2
Suecia
0.9
5.9
0.1
39.5
53.6
Finlandia
21.1
14.4
9.1
31.2
24.2
Islandia
0.1
94.5
5.4
Eslovenia
16.3
15.7
3.2
28.6
33.1
0.1
(%)
Como se puede observar en la tabla, las diferencias existentes entre los diferentes
pases son en algunos casos muy grandes. De este modo, este es un parmetro que no debe
ser obviado, puesto que afectar de manera decisiva al impacto ambiental y econmico de
un determinado producto.
129
Resultados y Discusin
Efecto invernadero. Esta es una de las categoras obligatorias en todo anlisis LCA,
y est relacionada directamente con la cantidad y tipo de gases que se desprenden
como subproductos del proceso que estamos siguiendo. En este caso, la unidad de
trabajo es el kg CO2 eq4, de manera que cada gas liberado al medio ambiente se
normaliza por conversin a estas unidades. As, el valor numrico correspondiente
al efecto invernadero se obtiene por suma de los kg de CO2 eq. que aporta cada uno
de los gases desechados a lo largo de todo el proceso (tanto en la propia etapa de
produccin como en la de obtencin de energa).
3. The Netherlands Guide on Environmental Life Cycle Assessment [en lnea],Jeroen Guine, Institute of Environmental
Sciences <http://www.leidenuniv.nl/cml/ssp/projects/lca2/index.html>, [5 Marzo 2004]
4. Climate Change, Radiative Forcing of Climate Change, Cambridge University Press, (1995)
5. Integrated Pollution Prevention and Control (IPPC); Reference Document on Best Available Techniques in the Chlor-Alkali
Manufacturing Industry (December 2001) [en lnea], European Comission, <www.envir.ee/ippc/docs/chlor-alkali.doc>, [5
Marzo 2004]
130
Resultados y Discusin
Tabla 3.13. Resultados para las categoras estudiadas en LCA simplificado de los procesos de
produccin de NaOH electroltico y de resina
Categoras
Efecto invernadero
(kg CO2 eq.)
Disminucin de las
reservas abiticas
(kg Sb eq.)
Uranio utilizado
(g U eq.)
Mercurio
liberado
(g Hg eq.)
Proceso electroltico
25.31
1172.2
13
0.4 - 6
Proceso de resina
275.45
2048.9
2.65
1.1
131
Resultados y Discusin
132
Conclusiones
1A
133
Conclusiones
1A
2B
3C
134
Conclusiones
4 CONCLUSIONES
A partir de los resultados descritos en la presente memoria, podemos resumir que la
contribucin de estos estudios tiene como principal objetivo el conocimiento del impacto
ambiental del mercurio y su posible reduccin. De esta forma, el contenido de esta tesis
est constituido por dos grandes bloques, la determinacin de especies de mercurio y el
diseo de un mtodo de sntesis de hidrxidos alcalinos alternativo al de ctodo de
mercurio. La utilizacin de la herramienta LCA para el anlisis y comparacin
medioambiental de dos procesos de sntesis nos proporciona una valiosa informacin para
determinar el alcance de estos resultados.
Refirindonos al primero de los bloques, la especiacin de mercurio, el trabajo se ha
basado principalmente en el desarrollo y mejora de metodologas para la determinacin de
dichas especies. An y siendo un clsico en el campo de la contaminacin medioambiental,
el estudio de la presencia de mercurio en diferentes mbitos de nuestro medio ambiente
sigue siendo de la mxima actualidad. La determinacin de diferentes especies de
mercurio sigue necesitando de nuevos avances y descubrimientos que permitan ante todo
mayor sensibilidad y especificidad en el anlisis. Desde este punto de vista, el trabajo
realizado en esta tesis utilizando metodologas clsicas creemos puede ayudar a la
mejora de tcnicas ya existentes. Como se ha podido observar, uno de los reactivos
tradicionalmente utilizados como agente extractante (la cisteina), puede colaborar en la
degradacin de especies organometlicas de mercurio, ayudado por variables como la
exposicin a la luz o tiempos de anlisis excesivamente largos. Adems, a nivel de qumica
elemental en el ciclo del mercurio, los mecanismos propuestos pueden ayudar no slo a la
comprensin de este caso concreto, sino que dada la importancia de la relacin Hg S, se
pueden establecer paralelismos claros con la interaccin Hg cidos hmicos o Hg
complejantes orgnicos.
El trabajo con el equipo de Electroforesis Capilar de Alta Resolucin permiti llegar a
resultados muy positivos en cuanto a la calidad del mtodo de preconcentracin
desarrollado. Queda fuera de toda duda la capacidad separativa de la tcnica, hecho que
le confiere un gran atractivo desde el punto de la especiacin qumica. Si adems
disponemos de herramientas para mejorar sus lmites de deteccin, est claro que estamos
ante una opcin de futuro indiscutible. En el trabajo presentado en esta tesis, se
consiguieron reducir los lmites de deteccin entre 5 y 10 veces respecto a aqullos
obtenidos sin metodologa de preconcentracin alguna. Ya se expuso en su momento que
stos eran todava lmites altos, y el trabajo en esa direccin es claro y definido. Pero lo
que quizs cabe remarcar es el hecho de haber presentado un mtodo estndar,
135
Conclusiones
compatible con todos los equipos presentes en el mercado, y que adems es barato y muy
eficiente.
El ltimo de los trabajos realizados a partir de mtodos clsicos fue la determinacin
directa de especies de mercurio en muestras acuosas. Esta metodologa supone un salto
cualitativo en el concepto de especiacin por la capacidad demostrada en la identificacin
de mercurio inorgnico y metilmercurio (en muestras acuosas) sin utilizar previamente una
tcnica separativa. Sin embargo, es cierto que la metodologa presentada todava tiene
ciertas limitaciones, como la necesidad de demostrar su aplicabilidad a muestras reales o
que los lmites de deteccin se encuentren todava lejos de los valores normales en aguas
contaminadas.
Sin duda uno de los temas clave presentado en esta tesis ha sido la determinacin de
especies de mercurio por tcnicas XAFS. Y lo es por su novedad, por su aplicabilidad en el
campo medioambiental y por la calidad de la informacin que nos puede aportar. De lo
presentado sobre XAFS en esta tesis, es importante destacar el avance que ha supuesto su
introduccin como tcnica para estudios medioambientales en nuestro grupo de
investigacin. Debe adems tenerse en cuenta que los presupuestos para la construccin
del primer sincrotrn espaol han sido ya aceptados por la Generalitat de Catalua y el
Gobierno Central conjuntamente, aumentando as la probabilidad de obtencin de tiempo
de haz y haciendo factible profundizar en las ideas aqu expuestas en un futuro a medio
plazo (inauguracin prevista para el ao 2008).
As, podemos afirmar que en esta tesis se han demostrado las posibilidades de la tcnica
en el caso concreto de la especiacin de mercurio. Se han puesto a punto las condiciones
de anlisis en una lnea de haz, y diversas especies de mercurio han podido ser
identificadas y cuantificadas. Es cierto que los lmites de deteccin que se han barajado
distan mucho de los que podramos denominar como reales, pero este ha sido slo un
primer paso, como demuestran los tres proyectos aceptados en el ao 2003 o los dos ya
aceptados para el ao 2004 y que buscan entre otras cosas la reduccin significativa de
esos lmites de deteccin. Otro problema al que se ha hecho frente ha sido la presencia de
ciertos elementos, que de encontrarse en la muestra, interfieren en la obtencin de
buenos resultados. As, arsnico y zinc dan lugar a interferencias importantes cuando se
encuentran en concentraciones elevadas. El problema es slo relativo, pues existen en el
mercado programas capaces de deconvolucionar picos superpuestos. En cuanto al problema
del hierro, y la posible saturacin del detector de fluorescencia, la utilizacin de filtros
aparece como una opcin factible y fcil.
Ya ha sido comentada la importancia en las metodologas XAFS del tratamiento de
datos. A partir de la comparacin realizada en esta tesis entre dos programas de
136
Conclusiones
Las condiciones finales elegidas fueron un compromiso entre las tres categoras.
Respecto a la regeneracin de la resina de intercambio inico y a pesar de los intentos
por utilizar como agente regenerante la salmorra residual de una planta desalinizadora,
esta opcin no pudo ser dada por buena por la presencia de elementos interferentes que
impedan la correcta regeneracin de la resina. As pues, se concluy que la resina deba
ser regenerada a partir de mezclas de MCl/M2SO4 (M: Na o K). Cabe destacar que los
137
Conclusiones
Abstract
This paper concerns with the evaluation of a speciation methodology for organomercury compounds in soil samples. The
methodology used is based on the West method, which includes an acid leaching of the soil followed by an organic solvent
extraction and a re-extraction with a selective aqueous reagent (cysteine). The separation and quantification of organomercury
compounds is achieved by capillary electrophoresis technique, using an ultraviolet-visible detection system (HPCE-UV). The
main steps in the speciation process are discussed, from the point of view of possible degradation processes. Frozen, de-frozen
and non-frozen soil samples have been analyzed, characterizing the influence of light and temperature. Equally, the effect
of light and temperature, as well as the chemical form of cysteine, have been analyzed on aqueous extracts. Organomercury
degradation has been observed in the aqueous phase and related results are discussed.
2003 Elsevier Science B.V. All rights reserved.
Keywords: Methylmercury; Phenylmercury; Cysteine; Capillary electrophoresis; Mercury species degradation
1. Introduction
Mercury is a very toxic element, which is liberated
to the environment in considerable amounts by anthropogenic activities [1]. Fossil fuels combustions,
calcinations of pyrites or releases from former industries lead to a liberation of thousands of tones of
different mercury pollutants to the biosphere [2]. Elemental and inorganic mercury are the most commons
forms in which mercury is released, but not the only
ones. Organomercury compounds are also generated
in some plastic producing industries, or used as pesticides because of their high toxicity. Methylmecury
is also biologically generated, under particular condi Corresponding author. Tel.: +34-935812903;
fax: +34-935811985.
E-mail address: manuel.valiente@uab.es (M. Valiente).
0003-2670/03/$ see front matter 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0003-2670(02)01657-4
220
2. Experimental
2.1. Reagents
Methylmercury chloride, phenylmercury chloride,
mercury chloride and l-cysteine hydrochloride were
of analytical grade from Alfa Aesar (Germany).
Anhydrous l-cysteine was of analytical grade from
Merck (Germany). Phenylmercury and methylmercury appear to have a very low solubility in water. In
consequence, 1 g/l stock solutions were prepared in
methanol analytical grade, purchased from Panreac
(Spain).
Hydrochloric acid 37% (low mercury content) from
J.T. Baker (USA) was used for the soil leaching process, while analytical grade toluene (Panreac) was
used as organic diluent in the extraction process.
Cysteine solutions were prepared in Milli-Q water, by dissolving the necessary cysteine compound to
achieve 0.01, 0.1 or 1% weight content.
Instra-analysed nitric acid 6970% (low mercury
content) from J.T. Baker was used for total mercury
determination by an Analytical Microwave system
(MARS-5 model from CEM Corporation, USA).
Standards and samples to be analyzed by ICP-OES
were prepared by using gold as stabilizing agent. The
content of gold (1 g/l stock of standard solution, from
Panreac) was two times the expected mercury content.
2.2. Methodology
Two different experimental series were performed
considering the study system described above. For the
soil samples stability measurements, artificial samples
were prepared by spiking clean samples with known
amounts of organomercury species. Considering these
samples, total mercury content was also followed by
ICP-OES, previous MW digestion.
Speciation experiments were done by an HPCE-UV
system. First sample series corresponded to the soil
stability experiments. Cysteine phase, from the soil
re-extraction step was in consequence analyzed. Second sample series corresponded to the study of the
cysteine phase stability. In order to demonstrate the
influence of difference parameters as temperature,
sunlight and cysteine type on mercury stability, three
different calibration standards were prepared and
kept under given conditions. The evolution of these
samples was followed by the HPCE-UV system.
2.2.1. Soil samples preparation
Spiked soil samples were obtained as described
by Vzquez et al. [55]. Soil samples were mixed
with methanol solutions containing the desired
221
Table 1
HPCE-UV conditions for the determination of organomercury compounds
Capillary characteristics
Buffer
Voltage
Analysis time
Injection volume
Detection
Capillary temperature
Flushing solutions
60 cm length (50 to the detection window) fused silica capillary, poliimine coating and 75 m i.d.
100 mM borate buffer, pH 8.30, 10% methanol content
22 kV, normal polarity
7 min, approximate electrophoretic times: HgCH3 + : 4.1 min; HgC6 H5 + : 4.3 min; cysteine: 4.9 min
and Hg2+ : 5.4 min
77.4 nl 0.8% total capillary replacing
UV detection, at 200 nm
303 K
NaOH 0.5 M, Milli-Q water and conditioning with borate buffer
222
Table 2
MW conditions for soil digestions (EPA Method 3051)
Sample weight
Acid
Temperature
Pressure
Table 3
Ageing effect for methylmercury and phenylmercury in a soil sample kept under different conditions
Day
Concentration (ppm)
S.D.
Concentration (ppm)
S.D.
Concentration (ppm)
S.D.
30.4
31.4
27.2
2.3
1.5
1.3
24.9
29.5
26.4
1.0
2.0
1.3
30.6
26.2
25.3
1.3
1.3
2.0
HgC6 H5 Cl
1
41.6
3
35.8
5
46.7
1.0
2.7
2.3
46.8
43.5
41.5
2.7
2.2
1.3
44.3
42.2
46.8
0.7
1.0
1.0
HgCH3 Cl
1
3
5
223
Fig. 1. Aging effect (in terms of total mercury content) for a soil sample spiked with 450 mg/l of methylmercury and 450 mg/l of
phenylmercury. Each point is the mean value of 3 replicates.
224
X. Gaona, M. Valiente / Analytica Chimica Acta 480 (2003) 219230
Fig. 2. Aging effect for sample 1, containing 5 mg/l of HgC6 H5 Cl. From day 1 to 8, plus eight spiked with 5 mg/l Hg2+ .
Fig. 3. Aging effect for sample 2, containing 5 mg/l of HgCH3 Cl and 5 mg/l of HgC6 H5 Cl. From day 1 to 8, plus eight spiked with 5 mg/l Hg2+ .
225
226
Fig. 4. HgCl2 calibration curves for an 8 days lifetime sample. A 0.1% anhydrous cysteine media, 277 K and light protection: (, ) day
1; (, - - -) day 4; (, ) day 8.
O2 +
HgCH3 +
Hg2+
HgC6 H5 +
(1)
(2)
Fig. 5. HgCl2 calibration curves for an 8 days lifetime sample. A 0.1% anhydrous cysteine media, room temperature and sunlight exposure:
(, ) day 1; (, - - -) day 4; (, ) day 8.
227
Fig. 6. HgCH3 Cl calibration curves for an 8 days lifetime sample. A 0.1% anhydrous cysteine media, room temperature and sunlight
exposure: (, ) day 1; () day 4; () day 8.
Fig. 7. HgCl2 calibration curves for an 8 days lifetime sample. A 0.1% anhydrous cysteine media, 277 K and light protection: (, ) day
1; (, - - -) day 4; (, ) day 8.
were present. However, Suda et al. [57] reported a further stabilization for mercury species when a cysteine
derivative (N-acetyl-l-cysteine) was used.
On the other hand, cysteine is easily dimerized to
cystine [5860] in a neutral/basic media, in presence
of dissolved oxygen.
(3)
228
Fig. 8. HgC6 H5 Cl calibration curves for an 8 days lifetime sample. A 0.1% anhydrous cysteine media, 277 K and light protection: (, )
day 1; (, - - -) day 4; (, ) day 8.
Fig. 9. tr(HgC6 H5 + )tr(HgCH3 + ) plot for cysteine hydrochloride solutions (circles) and anhydrous cysteine (triangles). Black: day 1;
white: day 4 and grey: day 8.
229
(4)
The proposed equation would explain the ambiguous
role of cysteine in an organomercury/cysteine system,
being able to stabilize mercury compounds when a
complex with them is formed, but acting as an initiator for the organomercury degradation when itself is
degraded to cystine.
The cysteine degradation has also an important effect on electrophoretical resolution. Fig. 9 shows the
migration time difference between methylmercury and
phenylmercury peaks.
We observe a clear resolution loss with cysteine
aging. Difference in electrophoretic time between
methylmercury and phenylmercury is reduced from
0.2 to 0.15 in 7 days.
Regarding the effect of the cysteine type on the
mercury degradation, no differences have been found
when using cysteine hydrochloride or anhydrous cysteine. Equally, the electrophoretical resolution is not
affected by the type of cysteine.
4. Conclusions
From the results reported here, we can conclude
that there is no degradation effect during a 5 days period on mercury compounds directly attached to soil,
either working at normal laboratory conditions (room
temperature and light exposure) or using frozen storing conditions.
Regarding the extracted aqueous phase stability,
some possible degradation factors must be taken
into account. Sunlight and cystine, a degradation
sub-product of cysteine, play an important role on the
degradation of organomercury species to inorganic
mercury. An important degradation has been demonstrated for a 4 days aging of the aqueous sample (it
is the minimum time to observe the presence of inorganic mercury using the CE technique), being that
degradation stronger when both factors (light and
cysteine) are present. No differences on stability have
been found when working with different forms of
cysteine (hydrochloride or anhydrous).
Degradation of cysteine appears to have also a considerable effect on the electrophoretic resolution. The
lost on resolution is about 25% for an 8 days aged
aqueous solution.
Acknowledgements
This work has been performed under the economical support of the European Commission (project
EVK1-CT-19999-00002). Xavier Gaona wants also
to acknowledge the economic support received from
Generalitat de Catalunya.
References
[1] J.M. Pacyna, J. Mnch, Water Air Soil Pollut. 56 (1991) 51
61.
[2] E.J. Zillioux, D.B. Porcella, J.M. Benoit, Environ. Toxicol.
Chem. 12 (1993) 22452264.
[3] F. Baldi, M. Filippelli, G.J. Olson, Microbial Ecol. 17 (1989)
263274.
[4] M.C. Marvin-DiPasquale, R.S. Oremland, Environ. Sci.
Technol. 32 (1998) 25562563.
[5] J.K. King, J.E. Kostka, M.E. Frischer, F.M. Saunders, R.A.
Jahnke, Environ. Sci. Technol. 35 (2001) 24912496.
[6] R.S. Oremland, C.W. Culbertson, M.R. Winfrey, Appl.
Environ. Microb. 57 (1991) 130137.
[7] P. Porvari, M. Verta, Water Air Soil Pollut. 80 (1995) 765
773.
[8] P. Quevauviller, I. Drabaek, H. Muntau, M. Bianchi, A.
Bortoli, B. Griepink, Trends Anal. Chem. 15 (1996) 390397.
[9] M.C.R. Rezende, R.C. Campos, A.J. Curtis, J. Anal. Atom.
Spectrom. 8 (1993) 247251.
[10] D. Purves, Fundamental aspects of pollution control and
environmental sciences, in: Trace Elements Contamination of
the Environment, vol. 7, Elsevier, New York, 1987, p. 17.
[11] J.R.D. Guimaraes, M. Meili, L.D. Hylander, E. de Castro e
Silva, M. Roulet, J.B. Narvaez, R. Alves de Lemos, Sci. Total
Environ. 261 (2000) 99107.
[12] J.R.D. Guimaraes, M. Roulet, M. Lucotte, D. Mergler, Sci.
Total Environ. 261 (2000) 9198.
[13] S. Heaven, M.A. Ilyushchenko, K. Kamberov, Sci. Total
Environ. 260 (2000) 4555.
[14] H. Biester, M. Gosar, G. Mller, J. Geochem. Explor. 65
(1999) 195204.
230
1. INTRODUCTION
During the last years, the popularity of XANES (X-ray Absorption Near Edge
Spectroscopy) has been notoriously increased. Several applications in the field of
chemical speciation have been found to be adequate for such kind of analysis. Its
capacity to distinguish similar chemical species with slight differences and its simple
sample pre-treatment makes it specially appropriate to environmental analysis, as the
direct speciation in polluted soils (Calmano et al., 2001; Welter et al., 1999; Welter et
al., 1998), the recognition of the bonding character of a contaminant in a soil (Skyllberg
et al., 2000), the pollutant speciation on fly ash (Osn et al., 1997), the air particulates
speciation (Ressler et al., 2000; Ressler et al., 1999), etc. XANES is not able to provide
information on either the distance between atoms or the number of neighbouring atoms
as it occurs with EXAFS, but XANES efficiency with fingerprint information is an
appreciated value.
The main drawback of XANES analysis is its high detection limits. Although important
improvements are continuously introduced, detections limits of hundreds of ppm are
quite normal.
As most of XAS techniques, data evaluation is the more complex part of XANES
analysis. There is a large list of computer programs (Catalogue of XAFS analysis
programs, 2001) for the data evaluation of experiments developed in the X-Ray
Absorption field. However, for XANES analysis, the number of programs is very
reduced. This paper concerns with the behaviour of two of these programs, WinXAS
and SpecManger, paying special attention to:
-
The ability of the program to deal with data from different sources and modes (i.e.,
transmittance and fluorescence), also being prepared to have a data input for an
external reference
The ability to work not only in the post-Edge region (usual in the EXAFS field) but
also in the pre-Edge region (needed for the treatment of XANES data)
being a well-tested program, with a complete manual and the experience of end-users
considered in the two versions appeared. On the other hand, this program has some
drawbacks, as the limited performance and exhaustivity when specific data, as the
XANES field, are treated.
Regarding SpecManger, it is a non-commercial program developed at the Technical
University of Hamburg-Harburg by Stefan Mangold. However, SpecMANGER is not a
complete program, but a MACRO of the commercial program IGOR Pro
(http://www.wavemetrics.com/). Because of that, the IGOR Pro license must be
acquired with a pricing from 395$ to 550$ (academic and standard use respectively).
Both programs work under Windows interface, although SpecMANGER is also able to
properly work under MAC systems. The system requirements are similar in both cases
(WinXAS and IGOR Pro as support of SpecMANGER): a Pentium (I) with 16-32
Mbytes RAM is enough to work with, although RAM problems can appear depending
on the number and size of the treated files.
than 5 pixels are present. In that case, SpecMANGER is not able (at the moment) to get
such a fluorescence data. This problem is not present when working in absorbance
mode, where the scheme of three different ionisation chambers is the usual one, with
practically no variation between different synchrotron facilities.
2m
(E E exp
0 )
2
h
(1)
Where m is the particle mass, h the reduced Planck constant, E the energy value and
E0exp the Edge energy value obtained experimentally.
Due to this, programs that are not able to perform the complete treatment in Energy
space cannot be considered for XANES analysis.
WinXAS
The energy correction is performed in WinXAS by using the reference spectrum. A
reference energy value is obtained from it, and compared with an energy value given by
the user. The difference between both values is then applied to the sample spectrum
corresponding to the reference spectrum analysed.
WinXAS offers three different ways to obtain the reference energy value:
-
The precision needed in this type of analysis automatically disables the first option
mentioned. The second and the third option can be equally used, but the last one is
recommended because of the easier operation. In this option, the energy is corrected by
using the first or the second derivative of a given area of the reference spectra, which is
selected by the user. Figure 1 shows the spectrum of a lead foil, obtained as a reference
of a lead polluted soil sample.
Figure 1. Section of Energy correction in WinXAS. Red: XANES spectrum, Blue: First derivative,
Green: Second derivative
Various energy values can be taken as reference. However, and due to signal stability,
the value of 0 for the second derivative, located at the inflexion point in the Edge jump,
is recommended.
SpecManger
For SpecManger, a comparison between the reference spectrum of the sample (or the
reference compound) and the series reference spectrum is performed. In this case, the
series reference spectrum corresponds to a reference spectrum (from the reference
metallic foil) obtained in the same beam time that the target spectrum.
This comparison is not limited to a single point but established in an energy range
selected by the user and which is normally located in the XANES region (from some eV
before the Edge to some eV after the Edge). The program obtains (by an optimisation of
the least squares) the energy displacement between the series reference spectrum and
the reference spectrum of the sample, and applies this difference to the sample
spectrum. In this way, differences between spectra are removed, not only for spectra
obtained at the same beam time but also for those spectra obtained at different beam
times.
(E) =
(E) 0 (E)
0 (E)
(2)
Due to the difficulty to the experimental determination of 0(E), the usual way to
correct the background is to adjust polynomial functions, which are later substracted, to
obtain the final corrected signal.
Regarding fluorescence experiments, the background correction is still more complex.
In this case, autoabsorption effects must be also taken into account.
WinXAS
The program offers three different possibilities for background removal:
-
Polynomial fitting
Vitoreen fitting
Polynomial and Vitoreen fittings are limited to the pre-Edge area. In consequence, the
option Adjustment by two polynomial must be chosen for XANES analysis. The degree
of polynomial used in each region (pre and post Edge) is chosen by the user.
However, Cubic Spline adjustments are recommended for the post-Edge corrections.
This later option consists on fitting a series of polynomial functions. Good corrections
are achieved by this way if an adequate number of knots (in consequence polynomial
functions) is used.
Cubic Spline facilities are also present in WinXAS, although only ready to work in the
k space (in consequence not able for XANES data). Some conversions are suggested
from WinXAS manual to carry out if Cubic Spline adjustments in E space have to be
performed, as XANES requires. However, such conversions were not found to be
appropriate for a normal process of data treatment.
No specific background corrections for fluorescence data are present in WinXAS
program.
SpecManger
There exist two different approaches depending on the mode in which the spectrum was
obtained (fluorescence or absorbance). In fact, the main difference between these two
approaches is a first data treatment step the in case of fluorescence mode. This step
includes the introduction of the dead time, the angle between the incident radiation and
the sample, and in consequence the correction of the possible auto-absorption
phenomena (closely related with the mentioned angle). These two final effects are very
significative in case of highly concentrated samples and their correction must be
performed in a carefully way. This first fluorescence correction step includes also the
normalisation, weighing and average of the different pixels from the fluorescence
detector. All these routines are automatic and non-submitted to subjective decisions of
the user.
After this first step, all correction procedures are equal for both fluorescence and
absorbance spectra, and they will be described in the same way.
The pre-Edge correction is based on a fitting process with a third degree polynomial.
After the fitting, the function is subtracted from the initial spectrum to obtain a new
corrected spectrum in the pre-Edge region.
The case of post-Edge correction is a bit more complex, existing two different options:
-
Cubic Spline adjust. This is a much more complex and complete post-Edge
adjustment option. It is based on the fitting of a function conformed by several
fourth degree polynomials. The user must define the starting and ending points, the
number of polynomial, and the Energy value that will be used to normalise the
complete spectrum. Afterwards, the program optimises the starting point by looking
for the combination having the lower initial slope. The Cubic Spline correction
option is then a really complete tool, very close to the total automatisation of the
process. However, some points must be controlled, as over-fittings by the use of an
excess of polynomial or problems with non-common XAFS spectra (see PbCl2 case,
Figure 2) where EXAFS oscillations can present some irregularities.
10
0.80
0.78
Signal
0.76
S pline 8 knots
0.74
0.72
0.70
13.0
13.1
13.2
13.3
13.4
13.5x10
Energy (eV)
Poly nom ial (13051 eV)
104
N ormalized signal
102
100
98
96
13.0
13.2
13.4
13.6
13.8x 10
Energy (eV)
Figure 2. PbCl2 post-Edge adjustments with SpecManger program. A. Different options to the
function adjustment. Note how polynomial adjustments offer more problems when complicate
spectra as PbCl2 are analysed (in this case, test with different energy values as a start of the
polynomial adjust). Comparison between a Cubic Spline with 8 knots (9 polynomial) and 16 knots
(17 polynomial) is also done. 16 knots Cubic Spline give as a result an important over-adjustment.
B. Note the high variation of the resultant wave after the post-Edge correction depending de
chosen function
The program also allows the user to automatically compare both fluorescence and
absorbance spectra. This option is especially useful to detect auto-absorption
phenomena in the fluorescence mode.
A drawback appears when using the table that contains some necessary parameters for
internal calculations, as the fluorescence energy, the value of incident energy where the
spectrum starts, the weight of the different pixels, etc. These parameters are only
11
defined for a few elements (Pb, Cr and Cu) and there is no possibility for the user to
create new element profiles. A possible solution to such limitation is to enter the new
parameters and save the experiment. After this, every time the experiment is open, the
new parameters are also loaded.
12
1. The number of reference compound spectra is limited to 10. For higher amounts of
reference spectra, an error message is given and the program directly switches off.
This is a paramount problem, because the majority of data treatments start with
more than 10 reference spectra.
2. The maximal number of consecutive iterations is 50. If the adjustment is not
achieved after these first 50 iterations, the process must be continued by using the
concentrations obtained from the last iteration as starting values.
Figure 3. XANES Linear Combination section on WinXAS. Corrected spectrum of a lead polluted
soil sample compared with the resultant wave of 5 lead reference compounds (PbCl2, Pb sorbed on
humic acid, PbCO3, basic PbCO3 and PbSO4)
SpecManger
After these modifications, the fitting Panel is able to recognise and to work with the
corrected spectrum obtained in the former step. Then, the sample spectrum must be
loaded, the energy range (where the fitting process must be carried out) selected, and
any kind of weighing procedure assigned (if needed). The preferences option allows the
user to decide if small displacements in the energy are permitted to optimise the results.
13
It is also possible to select the minimum number of expected components and the
minimum amount of a compound to be considered as a component of the sample. After
that, the user can proceed to the fingerprint adjustment. The result is a list of
components with the corresponding proportions and standard deviations. The total value
of 2 is also given for the fit process. An overview of XANES linear combination for
SpecManger program is shown in Figure 4.
Figure 4. View of the Linear combination panel of SpecManger program. The plot presents both
spectra, the one from the sample and the one from the linear combination of reference compounds
spectra. In that case, the composition was 43.2% PbHum (lead sorbed on humic acid, 39.8%
PbCO3, 13.5% PbAcetate and 3.5% PbMont (lead sorbed on Monmorillonite, a clay mineral)
14
Mi_1_4komp
0.29
0.27
0.19
0.24
No
No
Mi_2_4komp
0.32
0.3
No
No
0.21
0.17
Mi_4_4komp
0.35
0.35
0.2
0.1
No
No
The results obtained by using WinXAS program are detailed in Table 2, while concrete
conditions for each analysis are shown below.
A. Sample mi_1_4komp
Component: Adjust 1 Adjust 2 Adjust 3* Adjust 4* Adjust 5* Adjust 6** Adjust 7*** Adjust 8****
Basic PbCO3
0.45
0.43
0.24
0.25
0.24
0.24
0.21
0.21
PbCO3
0.27
0.30
0.34
0.32
0.34
0.34
0.35
0.34
PbCl2
No
0.10
0.18
0.16
0.18
0.18
0.18
0.19
PbHumi2
No
0.17
0.24
0.17
0.24
0.24
0.26
0.26
PbMont4
0.16
No
No
0.10
No
No
No
No
PbOxalate
0.12
No
No
No
No
No
No
No
N. Init. Comp.
10
4
4
6
10
4
4
4
B. Sample mi_2_4komp
Component: Adjust 1 Adjust 2 Adjust 3* Adjust 4*
Basic PbCO3
0.35
0.3
0.33
0.26
PbCO3
0.35
0.36
0.37
0.39
PbSO4
0.14
0.18
0.12
0.18
PbCl2
0.16
No
0.18
No
PbS
No
0.16
No
0.18
N. Init. Comp. 10 inorg.
4
8
4
15
C. Sample mi_4_4komp
Component: Adjust 1* Adjust 2* Adjust 3* Adjust 4**** Adjust 5_
Basic PbCO3
0.34
0.35
0.34
0.34
0.33
PbCO3
0.37
0.40
0.38
0.40
0.40
PbCl2
0.14
0.15
0.15
0.16
0.16
Pb_mont_3
0.14
No
0.08
No
No
Pb_humi_2
No
0.11
0.05
0.10
0.10
N. Init. Comp.
10
10
5
10
4
Table 2. Results obtained with WinXAS program for the three different real samples. Bold
columns mark results closer to the real concentrations
Several tests were performed, in order to improve the results. The impossibility to
perform the analysis with all the spectra at the same time was the main cause of error
found, although a chemical criteria was applied when selecting the spectra.
16
The similarity between spectra PbMont and PbHum (Figure 5) was also found to be an
important source of problems. WinXAS was not able to discern at all between these two
compounds. The strategy was then to choose the one with lower residual.
Despite the agreement of some final results, this strategy was considered of an
important source of error in the analysis process.
Some problems with PbCl2 adjustment were also found, since the first oscillation after
the Edge is very low and its interpretation by the program difficult.
N ormalised signal
104
102
100
98
96
13.0
13.2
13.4
13.6
13.8x 10
Energy (eV)
The data analysis with SpecManger was much more simple, with results detailed in
Table 3 and conditions described below.
A. Sample mi_1_4komp
Compound
PbCO3
PbCO3 basic
PbCl2
PbHum
PbSO4
PbS
PbMont
N. Init. Comp.
Adjust 1
0.3
0.21
0.11
0.15
No
No
0.2
All (17)
Adjust 2 *
0.35
0.23
0.15
0.27
No
No
No
All (17)
17
B. Sample mi_2_4komp
Compound
PbCO3
PbCO3 basic
PbCl2
PbHum
PbSO4
PbS
PbMont
N. Init. Comp.
Adjust 1
0.38
0.3
No
No
0.1
0.11
0.11
All (17)
Adjust 2*
0.36
0.28
No
No
0.19
0.17
No
All (17)
C. Sample mi_4_4komp
Compound
PbCO3
PbCO3 basic
PbCl2
PbHum
PbSO4
PbS
PbMont
N. Init. Comp.
Adjust 1
0.14
No
No
0.23
0.45
0.17
No
All (17)
Adjust 2*
0.39
0.35
0.14
0.12
No
No
No
All (17)
Table 2. Results obtained with SpecManger program for the three different real samples. Bold
columns mark results closer to the real concentrations
Regarding the conditions of the different analysis performed with SpecManger program,
these are:
Conditions:
- Pre-Edge adjustment between 12800 12925 eV
- Post Edge adjustment with Cubic Spline of 8 knots for all spectra (both reference
compounds and samples)
- Start of the Cubic Spline adjustments in the post-Edge region at the point 74 (13050 eV) for
all the compounds/samples (equivalent to the first EXAFS oscillation), except for the PbCl2
reference. In this case, the Cubic Spline was started in the second EXAFS oscillation (point
100, energy 13076 eV) because of the low amplitude of the first oscillation
- End of the Cubic Spline adjustment for the post-Edge region in the point 380 (13900 eV)
- Reference point at 245 (13327,2 eV) for the normalisation process
- Energy region used in the comparison of the reference spectra for the energy correction
process: 40 140 (13000 13125,7 eV)
- Energy region used in the fingerprint adjustment: 40 177 (13000 13184,9 eV)
Conditions*:
The same as Conditions, except:
- Post-Edge polynomial adjust (instead of Cubic Spline) for PbCl2 case (starting energy
13072 eV)
The second option for analysis conditions used, appear to be an excellent solution
considering both results and complexity of the analysis. As happened with WinXAS,
18
problems with PbCl2 have been found. In this case, a polynomial approach seems to be
a better solution than Cubic Spline adjustments.
5. CONCLUSIONS
Taking into account the observed characteristics, it can be concluded that SpecManger
results in more accurate calculations for XANES data evaluation than WinXAS.
WinXAS appears to be a good program for general XAS data evaluation, or, as it is
already mentioned in the WinXAS manual, it is meant to complement rather than
substitute more advanced XAS data analysis programs, as occurs in that case with
XANES.
WinXAS is mainly deficient when working in fluorescence mode, especially regarding
data acquisition and correction tools. Despite this, results obtained by WinXAS with
transmission data are very close to the ones obtained by SpecManger.
SpecManger appears to be a very precise program for XANES data evaluation. All main
sections of a program forXANES data evaluation (data input, energy correction,
background removal and fingerprint adjustments) are very complete and accurate.
Patience use of SpecManger program in data evaluation will offer to the end-user good
and correct results for XANES experiments.
However, SpecManger is an unfinished program, far of being a commercial program
(maybe because it was not its first objective). The drawbacks of handling a noncommercial program must be also taken into account.
5. ACKNOWLEDGEMENTS
This work has been performed under the economical support of the European
Commission and the Marie Curie Fellowship Association.
19
6. REFERENCES
Calmano, W., Mangold, S. & Welter, E. (2001). Fresenius' J. Anal. Chem. 371(6), 823830
Catalogue of XAFS analysis programs. (2001). http://cars9.uchicago.edu/IXScgi/XAFS_Programs
Osn, J., Trk, B., Trk, S. & Jones, K.W. (1997). X-Ray Spectrom. 26, 37-44
Ressler, T., Wong, J., Roos, J. & Smith, I.L. (2000). Environ. Sci. Technol. 34, 950-958
Ressler, T., Wong, J. & Roos, J. (1999). J. Synchrotron Rad. 6, 656-658
Ressler, T. (1998). J. Synchrotron Rad. 5, 118-122
Skyllberg, U., Xia, K., Bloom, P.R., Nater, E.A. & Bleam W.F. (2000). J.Environ.Qual.
29, 855-65
Welter, E., Calmano, W., Mangold, S. & Troger, L. (1999). Fresenius' J. Anal. Chem.
364(3), 238-244
Welter, E., Calmano, W. & Mangold, S. (1998). Contam. Soil '98, Proc. Int. FZK/TNO
Conf., 6th. 2, 847-848
20
Abstract. This manuscript describes the analysis of polluted soil samples containing
different species of mercury by using XAFS (X-ray Absorption Spectroscopy)
techniques. Such techniques allow the direct analysis of the sample, while having a very
powerful capacity for species recognition. Different aspects of this process are presented
and discussed, such as sample preparation, experimental conditions and development at
the synchrotron facility (HASYLAB in Hamburg, Germany) including the treatment of
complex data by using the computer program SpecMANGER. The homogeneity of
target samples measured by XAFS was characterized by SEM-EDX analysis.
Although XAFS cannot be considered yet an analytical tool, the results achieved in this
study were compared with those obtained with the same samples by using an
operational speciation methodology based on a wet extraction of the mercury
compounds from the polluted soil samples (West method [1])
Due to the high detection limits of XAFS technique, the comparison was done with real
soil samples spiked with different concentrations of HgCl2, HgCH3Cl and HgC6H5Cl. In
general, the results reveal a good agreement between both techniques, although the
possible formation of HgO is suggested by XAFS.
Corresponding author phone: +34 935812903; fax: +34 5811985; e-mail: Manuel.Valiente@uab.es
Introduction
Mercury is a very toxic element present in the environment in various chemical forms.
Despite the high toxicity of almost all mercury compounds a classification must be
considered regarding its mobility, bio-availability and capacity for bio-accumulation.
Organomercury compounds are the most toxic mercury species. In the environment they
appear in low proportions (0-5% with regard to inorganic and metallic species), and
result not only from anthropogenic sources (HgCH3X as a fungicide and HgC6H5X as
fungicide and bactericide) but mainly from natural microbial mechanisms [3].
Therefore, it is necessary to develop reliable methodologies for the mercury speciation
in several matrices, such as soils and sediments. Despite the existence of some
antecedents [4,5], West developed during the late 60s the most famous method for
the determination of methylmercury in environmental samples [1]. It was based on an
acidic leaching from the soil/sludge samples followed by an extraction into an organic
phase (benzene). The method included a purification step, consisting of a stripping of
the organomercury compounds from the benzene extract into an aqueous phase
containing cysteine, a very selective reagent toward mercury. Then, the organomercury
compounds were again re-extracted to benzene after acidification of the aqueous phase
and injected into an GC system. This procedure was the starting point for several
methodologies [6,7]. Nowadays, benzene has been substituted by toluene in almost all
procedures and also other stripping reagents containing sulphur are used instead of
cysteine, such as thiosulfate [8] and dithizone [9]. Several chromatographic alternatives
are applied for the separation, some of them by using the aqueous extract for the
separation and quantification of mercury species, i.e. HPLC [10,11] or HPCE [12].
Distillation [13-15] and supercritical fluid extraction [16] also appear to be good
separation techniques.
However, such techniques present some important drawbacks, like the possible
transformation of Hg compounds during the extraction process. Therefore, both
artificial methylmercury generation during extraction and distillation process [17-20]
and mercury degradation [21-24] have been observed and discussed. Conventional
procedures are also very limited regarding the number and type of species that can be
determined.
In this concern, XAFS techniques are able to play a very important role, due to its high
speciation capacity with almost no sample pre-treatment. Thus, better information about
toxicity and mobility can be obtained, since both parameters appear to be closely related
to the species present. However, XAFS techniques cannot compete yet, from an
analytical point of view, with such conventional procedures, although they must be
considered a very powerful tool for future applicacions.
X-ray Absorption Spectroscopy (XAFS) [25-27]
XAFS techniques are based on the interaction of highly brilliant X-radiation from a
special X-ray source (synchrotron) with a sample containing the target element. The
spectra are taken over a given energy range close to the region of an electronic level of
the target element. At the same time, the absorbance (or the fluorescence) of the Xradiation by the sample is observed. When the X-rays reach the energy of the target
electronic level, an absorption edge is observed. In the proximity of this absorption edge
some interferences appear due to the interaction of the X-rays with the neighbouring
atoms of the target element. Such interferences provide important information about the
chemical environment of the latter. This high speciation capacity and the possibility of
measuring with almost no sample pre-treatment (and in consequence with a minimum
of sample modification), make XAFS techniques a very useful tool for environmental
studies.
The core of the technique is the construction of a complete database of spectra from
pure species of the target element (Hg). Then, the spectrum of the sample is compared
with such compounds and a model based on the linear combination of these spectra
which presents the best fitting to the sample spectrum is searched.
However, after the acquisition of data, a complex treatment of these data must be
performed in order to obtain valuable results. Such data treatment becomes specially
complex when working in fluorescence mode, where only a few XAFS programs [28]
are available. Both the complex data treatment and the high detection limits are the
main drawbacks of the technique.
Several applications for the XAFS techniques have been reported during the last decade
regarding environmental problems. Lead compounds in soils have been the object of
several studies by different research groups [29-33], while some publications deal with
adsorption phenomena of lead on different surfaces [34-36]. Because of the high
variation in the toxicity of the different chemical species, chromium [37-45] and arsenic
[46-52] have also been studied by XAFS techniques (which is easily able to distinguish
between oxidation states). Regarding mercury speciation, the number of publications is
lower, and the main part deals with sulphur-Hg bonding [53-55], by following S as the
target atom. However, Kim et al. [56,57] studied mercury compounds in a mine tailing
in California (USA) by following the Hg LIII absorption line, being able to differentiate
several mercury minerals.
Taking into account such considerations, the present work aims to characterize the
chemical speciation of mercury in spiked soil samples by using XAFS techniques. The
same samples will be analysed by extraction techniques in order to check the reliability
of XAFS data, while looking also at the possible advantages that XAFS techniques can
offer.
Experimental
Three samples were prepared by spiking a clean soil (without mercury) with mercury
compounds at different concentrations of HgCH3Cl, HgC6H5Cl and HgCl2. The clean
soil was taken from an intensive sampling performed at a former fertilizers industry.
Additionally, three samples of model soil components (Hg(NO3)2 adsorbed on
montmorillonite, kaolin and humic acid) were prepared and included in the reference
compounds database.
Sample preparation
Samples were classified in two different groups:
a) Soil samples were spiked with various mercury compounds (as described by
Vzquez et al. [58]) in order to test the possibilities of the XAFS technique for
mercury speciation. Shortly, these samples were milled and sieved to get particle
sizes under 100 m. Three grams of each sample were mixed with 20 mL methanol
solution containing the selected organomercury compounds. After three hours of
equilibration, samples were air dried in a fume hood (with light protection) until
constant weight. The obtained material was distributed in aliquots to perform
different analytical tests: 1.5 g for capillary electrophoresis (HPCE) analysis, 0.5 g
for total mercury analysis by ICP-OES after analytical microwave digestion and 1 g
for pellet preparation for XAFS analysis. The prepared samples and their
concentrations are shown in Table 1:
Table 1 Concentration of HgCH3+, inorganic mercury and HgC6H5+ in samples H1-H3. Quantification by
extraction and HPCE separation-detection. Each concentration means of three replicates
Sample
Amount
HgCH3+
Inorg. Hg
HgC6H5+
Conc.
Percent.
Conc.
Percent.
Conc.
Percent.
(gg1)
(%)
(gg-1)
(%)
(gg-1)
(%)
H1
3g
1380 100
43.9 2.8
1760 120
56.1 3.4
H2
3g
1440 150
35.8 3.2
2580 160
64.2 3.3
H3
3g
710 120
27.6 3.7
1260 150
48.8 5.0
610 100
23.7 3.1
Percent (%). Percentage of the specie over the total Hg content in the sample
H1 and H2 were used to test the ability of XAFS to distinguish between slight
concentration changes in a binary system. H3 was prepared to test the possibility to
differentiate between two organomercury compounds
b) Soil materials (clay minerals, humic acids, etc) play a very important roll in the
mercury speciation and behaviour in natural environments. Because of that,
Hg(NO3)2 was absorbed on three different soil materials (montmorillonite, kaolin
and humic acid, purchased from Merck, Germany) which were included as reference
compounds in our database. Three grams of each soil component were mixed with
20 mL of an aqueous solution of 375 mgL-1 Hg(NO3)2, resulting in a theoretical soil
concentration of 2500 gg-1. All the samples were protected against light during the
complete process. The pH was adjusted to 5 (except for the humic acid). The
achievement of the equilibrium was considered when pH remained constant. After
equilibration, samples were filtered through 0.45 m membranes to remove the
excess of water and non adsorbed mercury, and air dried in a fume hood (always
with light protection). The final concentration of each soil sample was determined
by ICP-OES analysis after microwave digestion (see table 2).
Table 2. Total Hg concentration in three spiked soil materials (montmorillonite, kaolin and humic acid).
Each concentration means of three replicates
Sample
Weight
pH
H8. Montmorillonite
3g
1640 110
5.07
H10. Kaolin
3g
1920 90
5.29
3g
2200 100
9.65
concentration of the organomercury compounds from the total value determined with
the digestion procedure.
Table 3. HPCE-UV conditions for the determination of organomercury compounds
Capillary characteristics
Buffer
Voltage
Analysis time
Injection volume
Detection
UV detection, at 200 nm
Capillary temperature
303K
Flushing solutions
XAFS measurements
Experiments were performed at the HASYLAB synchrotron facility in Hamburg
(Germany) at the A1 beamline [62]. The beamtime for the experiments lasted 6 days.
Beam details are shown in Table 4.
Table 4. Beam characteristics at HASYLAB A1 beamline
Source (4.5 GeV)
Monochromator
Si (1,1,1)
Flux at sample
Transmitance detection
Fluorescence detection
4 elements Ge-detector
In addition to soil samples also pellets of the reference compounds HgCl2, HgCH3Cl,
HgC6H5Cl, red HgO, HgSO4, Hg(Acetate)2, red HgS and black HgS were produced. All
pure Hg compounds were purchased from Sigma-Aldrich (Germany), and used as
received. Pellets were obtained by pressing at 5 tons cm-2 for 5 minutes a mixture of soil
8
sample (or pure Hg salt) with a diluting agent (polyethylene). The total amount of
sample in each pellet varied from 50 to 100 mg (30 to 60 mg in case of reference
compounds) and polyethylene from 100 to 150 mg (See Table 5).
Table 5. Details of the prepared pellets for XAFS analysis
Reference Compound /
Weight Polyethylene
Pellet Thickness
Sample
mg
Mg
mm
HgCl2
39
74
0.70
HgCH3Cl
40
102
0.88
HgC6H5Cl
57
107
1.01
HgO
59
100
0.87
HgSO4
53
106
0.82
Hg(Acetate)2
56
116
1.14
Red HgS
58
103
0.83
Black HgS
46
106
0.87
131
138
1.73
86
125
1.49
111
162
1.79
H1
51
100
1.03
H2
73
122
1.21
H3
72
121
1.19
TP1
140
167
2.79
Soil samples were analysed using the fluorescence mode, by setting the distance
between the sample holder and the fluorescence detector window to 0.5 cm, while
reference compounds were analysed in transmittance mode (concentration in pellet is
much higher and then it is easier the data treatment). The angle between the beam and
the sample was set at 45.
The photon absorption of the target element (Hg) was recorded at the edge energy for its
LIII line at 12284 eV. Considering LIII absorption line, the main fluorescence lines are:
LI=9988.8eV and LII=9897.6eV.
To compare and correct any energy displacement, a thin HgCl2 pellet was equally
measured as a reference for all samples.
On the other hand, since each spectrum region does not contain the same information,
an energy program for measurements was prepared (see Table 6). Within this program,
the measurement took more time for the edge jump and neighbouring area.
Table 6. Energy program used for XAFS measurements
Energy range (eV)
E (eV)
12000-12270
Fluorescence: 4; Transmitance 1
12270-12350
Fluorescence: 4; Transmitance 1
12350-12500
Fluorescence: 4; Transmitance 1
12500-12900
Fluorescence: 4; Transmitance 1
Data were acquired and treated by XANES (X-ray Absorption Near Edge
Spectroscopy). Such technique studies the region around the edge which appears to be
very efficient for speciation studies.
The computer program SpecManger developed by S.Mangold, was selected for data
treatment [63]. It includes the following steps:
-
Correction of the fluorescence spectra, i.e. correction of the deadtime, selfabsorption effect, weighing and average of the different fluorescence spectra
corresponding to the different pixels from the detector. SpecManger allows
introducing the deadtime (already known at the beamline), the angle sample-beam
(which must be taken into account for the self-absorption correction) and the
number of pixels of the fluorescence detector (which usually vary from facility to
facility). The self-absorption effect is calculated by extrapolating the expected
absorbance of the sample at the LIII fluorescence lines (LI=9988.8eV and
LII=9897.6eV). Such extrapolation is done from the XAFS absorption spectrum.
10
Energy correction. Each analysis provides two different spectra: the sample
spectrum and the reference compound spectrum (HgCl2 pellet). The reference
compound spectrum remains always constant allowing the comparison of sample
spectra and the correction of some possible energy deviations. Usually, commercial
XANES programs perform this correction by comparing one point of the reference
spectra, while SpecManger works with a multipoint spectra comparison.
SEM-EDX analysis
Pellets prepared for XAFS experiments were also analysed by SEM-EDX, looking for
sample heterogeneities.
In order to achieve adequate conductivity in the soil/polyethylene samples, pellets were
submitted to a carbon coating treatment.
11
12
Figure 2. SEM picture (a); X-ray fluorescence spectra of Hg(LI) (b) and Hg(M) (c) for a mercury-rich
crystal from sample H1
XAFS measurements
Figure 3 shows corrected fluorescence spectra for all measured reference compounds
while Tables 7 and 8 show the composition of the samples H1 and H2 respectively,
compared with data from operational speciation (extraction + HPCE analysis).
13
250
Hg(NO3)2 adsorbed on hum.acid
Red HgS
200
Hg(NO3)2 adsorbed on kaolin
Hg(NO3)2 adsorbed on montm.
Black HgS
150
Hg(Acetate)2
HgSO4
HgO
100
HgPhCl
HgMeCl
HgCl2
50
0
12200
12250
12300
12350
12400
12450
12500
Energy (e V)
Figure 3. Corrected fluorescence spectra of: HgCl2, HgCH3Cl, HgC6H5Cl, HgO, HgSO4, red HgS, black HgS
and Hg(Acetate)2, Hg adsorbed on kaolin, Hg adsorbed on montmorillonite and Hg adsorbed on humic acid.
Each spectrum mean of 4 spectra. Shifts of 50, 100 and 150 units applied to some spectra in the Normalized
Fluorescence axis.
Table 5. H1 composition obtained from SpecManger compared with an operational speciation procedure
(extraction + HPCE analysis). % Percentage of the specie over the total Hg content in the sample. Error for
XAFS results expressed as the standard deviation from a non-linear fitting. Error for extraction expressed as
the confidence interval from a three replicates measurement
Compound
HgCl2
50 6
56.1 3.4
HgO
83
HgCH3Cl
42 6
43.9 2.8
14
Table 6. H2 composition obtained from SpecManger compared with an operational speciation procedure
(extraction + HPCE analysis). % Percentage of the specie over the total Hg content in the sample. Error for
XAFS results expressed as the standard deviation from a non-linear fitting. Error for extraction expressed as
the confidence interval from a three replicates measurement
Compound
HgCl2
60 7
64.2 3.3
HgO
42
HgCH3Cl
38 7
36 3.2
All reference spectra showed remarkable differences in the XANES region, enabling the
determination of single and mixed compounds in soil samples (see Figure 3). However,
it must be considered that the less polluted sample quantified had a total Hg content of
1600 gg-1. Samples with much lower Hg content are the object of future work. In
these cases, it is always necessary to deal with three factors:
-
The higher the time of signal accumulation, the higher is the sensitivity of
fluorescence detectors. However, this leads to longer analysis times.
The intensity of the beam and the matrix of the sample. The higher the intensity of
the beam, the higher the fluorescence signal is.
The presence of interferent elements. Interferences can affect the data quality in two
different manners:
40 00
Zn
Fe
As
30 00
20 00
Cu
10 00
Zn
Hg
0
3 40
44 0
540
64 0
740
C h an n el
Figure 4. Complete fluorescence window for sample TP1. Energy of the inciden X-radiation: 12300eV
Table 7. Elemental composition of the sample TP1. Each concentration value averages three replicates
Element
Concentration (g/Kg)
Hg
0.56 0.09
As
4.6 0.9
Zn
4.3 0.1
Pb
22.5 0.4
Cu
2.1 0.1
Fe
43.1 0.7
16
Considering qualitative data, XAFS is able to obtain information on the bonding type of
the target analyte. As described elsewhere [64,65] the distance between the two
absorption maxima in the XAFS spectra (see Figure 5a) is related to the covalent or
ionic character of the Hg-Ligand bonding (Ligands: Cl, O, CH3, C6H5, Acetate, etc). In
general, the higher the distance, the lower is the covalent character, as it has been
demonstrated elsewhere for a Cl-Br-I series [52].
The determination of the energy difference between both edge steps was performed
from the first derivative of the original spectra (see Figure 5b), where the inflexion
points in the original spectra appear as maxima and energy differences are measured
much more easily. Table 8 shows the energy differences found for HgCl2, HgCH3Cl and
HgC6H5Cl reference compounds and the spiked samples H1, H2 and H3.
17
60
40
HgCl_Fluo
HgMeCl_Fluo
HgPh_Fluo
20
First D erivative
10
80
0
12200
12
100
12250
12300
12350
E nergy ( eV)
12400
8
6
4
2
0
b
12200
12300
Energy (eV )
12400
Figure 5. a. Corrected fluorescence spectra of: HgCl2, HgCH3Cl and HgC6H5Cl; b. First derivative of the
original spectra
Table 8. Energy differences between two absorption maxima in Hg absorption edge (from first derivative
spectra of both reference compounds and spiked samples)
Reference Compound / Sample
E (eV)
HgCl2
8.8
HgCH3Cl
8.5
HgC6H5Cl
8.5
8.6
8.2
The energy differences obtained from reference compounds follow the expected
behaviour with respect to the ionic character of the Hg-Ligand bonding: HgCl2 >
HgCH3Cl > HgC6H5Cl. The much lower value found for phenylmercury chloride can be
explained in terms of electron de-localisation in between the mercury atom and the
aromatic ring (in consequence higher covalent bonding character).
The spiked samples (H1, H2 and H3) show similar tendencies as the reference
compounds.
18
Conclusions
Our studies represent a first step for deeper experiments in the field of mercury
speciation by XAFS techniques.
The spectra of the key species HgCl2, HgCH3Cl and HgC6H5Cl differ enough to allow
the correct identification if valuable data have been obtained. In fact, a good correlation
with conventional extraction results has been achieved. This study has also
demonstrated the possibility to distinguish between other mercury compounds such as
HgO, red HgS, black HgS, Hg(acetate)2, HgSO4 or mercury adsorbed in different soil
materials such as humic acids or clay minerals (montmorillonite and kaolin), since all of
them have also shown enough differences (see Figure 3), in agreement with other
publications [57]. Considering qualitative data, the compounds HgCl2, HgCH3Cl and
HgC6H5Cl have been succesfully classified by taking into account ionic character
bonding.
For the samples analysed, there has been no problem regarding the detection limits of
the technique. The lowest concentration determined was about 1600 gg-1 of mercury
in a soil sample, and four replicates were enough to offer a good result working in
fluorescence mode. The response was worst when working in the less sensitive
transmittance mode.
Despite some irregularities found in the pressed pellets by SEM-EDX techniques, the
homogenisation of the samples was found to be enough in order to obtain reproducible
XAFS results.
19
Future work
Once demonstrated the possibility to apply XAFS techniques to mercury speciation, the
main goal will concern the application of these techniques to real environmental
samples. That will mainly refer to the improvement of the detection limits. Two
different strategies will be followed to achieve such objective:
-
The use of more brilliant sources and more sensible detectors. However, under such
condictions it will be also necessary to deal with fluorescence coming from other
elements present in the sample.
Bringing XANES to the micro scale level, the so called -XANES [66]. This
technique combines the SEM-EDX imaging capacity with the XANES speciation
power, and it is performed on special beamlines with beam-focusing tools. Thus,
much narrower beams are obtained and more specific areas on the sample can be
analysed. The trick consists of identifying first by SEM-EDX the presence of Hgrich particles in the sample to allow focusing the beam to such areas. The main
drawback of this strategy is the need to deal with sample heterogeneities. However,
important advantages are also obtained: the Hg concentration in the particles is
much higher compared to the whole sample, and the presence of paralel elements is
partially removed, as the beam is focused in Hg-rich particles.
Acknowledgements
We gratefully acknowledge the scientifical and technical support of Edmund Welter
during synchrotron experiments at HASYLAB.
This work has been performed under the economical support of the European
Commission project EVK1-CT-1999-0002. Synchrotron experiments were supported by
the IHP-Contract HPRI-CT-1999-00040 of the European Commission.
20
XG wants to acknowledge the Marie Curie Fellowship Training program for being the
starting point for this interdisciplinary experience in-between classical and new
speciation technologies.
Catherine Krebs is gratefully acknowledged for her kind support to the manuscript
preparation.
References
1. West G (1968) Acta Chem Scand 22:2277-2280
2. West G (1968) Acta Chem Scand 22:2277-2280
3. Baldi F, Filippelli M, Olson GJ (1989) Microbial Ecol 17:263-274
4. Kitamura S, Tsukamura T, Hayakawa K, Sumino K, Shibata T (1966) Med Biol
72:274-280 in Japanesse
5. Sumino K (1968) Kobe J Med Sci 14:115-130
6. Snchez JE, Sanz-Medel A (1998) Talanta 47:509-524
7. Cela R, Lorenzo RA, Rubi E, Botana A, Valio M, Casais MC, Garca MS, Mejuto
MC, Bollan MH (1992) Environ Technol 13:11-22
8. Alli A, Jaff R, Jones RJ (1994) High Resol Chromatogr 17:745-748
9. Hardy S, Jones P (1997) J Chromatogr A 791:333-338
10. Hempel M, Hintelmann H, Wilken RD (1992) Analyst 117:669-672
11. Harrington CF (2000) TRAC-Trend Anal Chem 19:167-179
12. Jones P, Hardy S (1997) J Chromatogr A 765:345-352
13. Horvat M, Bloom NS, Liang L (1993) Anal Chim Acta 281:135-152
14. Eiden R, Falter R, Agustin-Castro B, Schler HF (1997) Fresen J Anal Chem
357:439-441
15. Horvat M, Liang L, Bloom NS (1993) Anal Chim Acta 282:153-168
21
30. Ostergren JD, Brown Jr GE, Parks GA, Tingle TN (1999) Environ Sci Technol
33:1627-1636
31. Welter E, Calmano W, Mangold S, Troger L (1999) Fresen J Anal Chem 364:238244
32. ODay PA, Carroll SA, Waychunas GA (1998) Environ Sci Technol 32:943-955
33. Manceau A, Boisset MC, Sarret G, Hazemann JL, Mench M, Cambier P, Prost R
(1996) Environ Sci Technol 30:1540-1552
34. Morin G, Ostergren JD, Juillot F, Ildefonse P, Calas G, Brown Jr GE (1999) Am
Mineral 84:420-434
35. Xia K, Bleam W, Helmke PA (1997) Geochim Cosmochim Ac 61:2211-2221
36. Brown Jr GE, Foster AL, Ostergren JD (1999) Proc Natl Acad Sci 96:3388-3395
37. Osn J, Trk B, Trk S, Jones KW (1997) X-Ray Spectrom 26:37-44
38. Thornton EC, Amonette JE (1999) Environ Sci Technol 33:4096-4101
39. Shaffer RE, Cross JO, Rose-Pehrsson SL, Elam WT (2001) Anal Chim Acta
442:295-304
40. Tokunaga TK, Wan J, Firestone MK, Hazen TC, Schwartz E, Sutton SR, Newville
M (2001) Environ Sci Technol 35:3169-3174
41. Taylor RW, Shen S, Bleam WF, Tu Shu-I (2000) Clay Clay Miner 48:648-654
42. Szulczewski MD, Helmke PA, Bleam WF (2001) Environ Sci Technol 35:11341141
43. Peterson ML, Brown Jr GE, Parks GA, Stein CL (1997) Geochim Cosmochim Ac
61:3399-3412
44. Szulczewski MD, Helmke PA, Bleam WF (1997) Environ Sci Technol 31:29542959
23
45. Hunter DB, Bertsch PM, Kemner KM, Clark SB (1997) J Phys IV 7:(C2, X-Ray
Absorption Fine Structure, Vol. 2) 767-771
46. Schwenzer SP, Tommaseo CE, Kersten M, Kirnbauer T (2001) Fresen J Anal Chem
371:927-933
47. Tahahashi Y et al. (2004) Environ Sci Technol 38:1038-1044
48. La Force MJ, Hansel CM, Fendorf S (2000) Environ Sci Technol 34:3937-3943
49. Reynolds JG, Naylor DV, Fendorf SE (1999) Soil Sci Soc Am J 63:1149-1156
50. Foster AL, Brown Jr GE, Tingle T, Parks GA (1998) Am Mineral 83:553-568
51. Foster AL, Brown Jr GE, Parks GA, Tingle TN, Voigt DE, Brantley SL (1997) J
Phys IV 7:(C2, X-Ray Absorption Fine Structure, Vol. 2)815-816
52. Hunter DB, Bertsch PM, Kemner KM, Clark SB (1997) J Phys IV 7:(C2, X-Ray
Absorption Fine Structure, Vol. 2)767-771
53. Hesterberg D, Chou JW, Hutchison KJ, Sayers DE (2001) Environ Sci Technol
35:2741-2745
54. Qian J, Skyllberg U, Tu Q, Bleam WF, Frech W (2000) Fresen J Anal Chem
367:467-473
55. Skyllberg U, Xia K, Bloom PR, Nater EA, Bleam WF (2000) J Environ Qual
29:855-865
56. Kim CS, Rytuba JJ, Brown Jr GE (1999) J Synchrotron Rad 6:648-650
57. Kim CS, Brown Jr GE, Rytuba JJ (2000) Sci Total Environ 261:157-168
58. Vzquez MJ, Carro AM, Lorenzo RA, Cela R (1997) Anal Chem 69:221-225
59. Lorenzo RA, Vzquez MJ, Carro AM, Cela R (1999) TRAC-Trend Anal Chem
18:410-416
60. Carro AM, Lorenzo RA, Cela R (1996) J Chromatogr A 730:345-351
61. Carro AM, Lorenzo RA, Cela R (1996) Mikrochim Acta 123:73-86
24
25
This paper presents the modelisation and optimisation procedure for a synthetic process
of tree different alkaline hydroxides: LiOH, NaOH and KOH. An ion exchange weak
acidic resin in M-form (M: Li, Na or K depending on the expected final product) was
used to obtain the correspondent hydroxide from lime milk. The regeneration of the
resin was achieved with an MCl/M2SO4 mixture.
Within these studies, it has been studied and optimised the influence of several
parameters on the final product concentration and purity, as the m/V ratio (m: mass of
resin, V: total solution volume), the lime milk concentration, the effect of the Ca(OH)2
particle size on the Kps and in consequence, on the final product purity. At the same
time, it has been modelled and successfully predicted the dependence of the final MOH
concentration on the separation factor (), taking into account the influence of other
parameters as the temperature, the m/V ratio or the Kps(Ca(OH)2). By using this
modelisation scheme, it has been also possible to estimate values for low Ca
concentrations in solution (XCa), hardly obtained by experimental procedures because of
the large time needed for resin equilibration.
Keywords: alkaline hydroxides, ion exchange, resin, modelisation, optimisation
E-mail: Manuel.Valiente@uab.es
Fax: ++34 935811985
Phone: ++34 935812903
Introduction
The alkaline hydroxides, and mainly sodium hydroxide, are very common reactive in
the actual chemical industry1. Sodium hydroxide is used for basification in a large list of
processes, due to its strong base character that allows its reaction with several metals,
i.e. aluminium or zinc. In addition, several of the sodium products obtained from the
hydroxide are water-soluble. Sodium hydroxide is used for example in paper and textile
industries, in several pharmaceutics processes, even though in the petroleum field2,3.
Despite being less important than sodium hydroxide, potassium hydroxide is also used
in huge amounts in the chemical industry, mainly as a precursor of other potassium
products. In less proportion, it is used in the tinctures production, for special soaps
manufacturing, and as a component in some glass types4.
The consumption of lithium hydroxide by chemical industries is much smaller than its
alkaline neighbours. It is an important component in the production of fats. In fact, in
the USA, 60% of the fats come from lithium soaps.
The way to produce such reagents depends on the hydroxide type. Both sodium and
potassium hydroxides are obtained by electrolytic means while lithium hydroxide is
produced from the reaction between the correspondent carbonate and the calcium
hydroxide.
Considering the electrolytic processes for these hydroxides, there exist three different
types which differs basically on the cathode used, and with different engineering
requirements, talking about reagents purities, energy consumption or environmental
relationship:
-
Mercury cathode process. It uses a cathode of mercury. Due to the high over
potential for the reduction of H+ to H2, the reduction reaction is Na+ to Na, thanks
also to the high stable amalgam NaHgx formed. Then, the amalgam reacts with
water and the desired product is obtained as:
Eq. 1
Diaphragm process. It uses a steel cathode. Differing with the mercury cathode
process, the reaction that takes place is the reduction of water to H2. In this process,
the cathode cell and the anode cell are separated by an asbestos barrier that allows
the free circulation of dissolved ions but not the gas bubbles formed in each cell.
With the diaphragm process the requirements considering both energy consumption
and purity of the initial reactive are less strict than the correspondent in the mercury
cathode process. On the other hand, the final product is less concentrated and pure.
As in the mercury cathode process, the main drawback is related to the environment,
because of the use of asbestos12 as a component of the diaphragm. However, the use
of this process is decreasing rapidly.
solubility product is strongly dependent on the particle size, this parameter has been
taken into account for the modelisation approach too.
It will be also introduced the ability to estimate (coefficient of selectivity) values for
Ca/Na and Ca/K systems, when working at low Ca concentrations. The knowledge of
this parameter can become of a paramount importance in this kind of processes, since
the affinity of the resin towards the different ions will determine both purity and
concentration of the final product.
Experimental Section
Materials, Apparatus and Analytical Methods
The study was performed with a weak carboxylic resin CNP 80, kindly supplied by
Bayer Hispania Industrial S.A. The ion exchange capacity of the resin was measured as
described elsewhere18, giving a result strongly dependent on the target ion: 9.1 0.2
meq Li+/g dry resin, 7.9 0.2 meq Na+/g dry resin and 7.0 0.1 meq K+/g dry resin.
This tendency was consistent with the capacity of the resin in the H+ form, 9.60.2.
All reagents were of p.a. grade. Sodium, potassium and lithium chloride, sodium
sulphate and hydrochloric acid were purchased from Panreac (Spain), while calcium
hydroxide was purchased from Probus (Spain).
Ion concentration in aqueous leachates was measured with an ICP-OES instrument,
ARL model 3410 spectrometer Fisons (USA), equipped with minitorch. Spectroscopic
specifications are shown in Table 1.
Li
Na
Ca
Emission line
670.78
589.59
766.49
422.67
0.0018
0.069
0.060
0.010
Interferences
V, Ti
Te, Ti, V
Ti
Fe
For the study of the solubility product (Kps) dependence with the m/V ratio, the jacked
cell for a batch system was used. Because of the experimental problems found at higher
temperatures, the study was performed only at 293K, while values at higher
temperatures were obtained from extrapolation thanks to tabulated data20.
Firstly, the synthetic process was done by mixing the necessary amount of Ca(OH)2 to
achieve a total content of 5, 10 or 15% (in weight) with the resin loaded with the target
ion (Li, Na or K depending on the study). The amount of resin varies also, depending on
the expected m/V (from 0.05 to 0.25). After reaching the equilibrium, a portion of the
solution was filtered and measured with the ICP-OES equipment, looking for M+ (M:
Li, Na or K) and Ca2+. The pH of the solution was also measured with a pHmeter Crison
507 (Spain). By having both Ca2+ and OH- concentrations, Kps was calculated as:
2
K ps = C Ca C OH
Eq. 2
Where CCa and COH are the concentrations in molarity of calcium and hydroxyl ion
respectively.
At the same time, an aliquot of the solid Ca(OH)2 excess was taken, in order to visually
determine the particle size by SEM with an electronic microscope equipment JEOL
JSM-6300 (Japan) equipped with an energy disperse spectrometer (EDX) LINK ISIS2000 from Oxford Instruments (Bucks, England).
Results and Discussion
Solubility product dependence with m/V. Effect of the particle size
The results corresponding to the experiments done with different m/V ratio at 293K are
shown in Figure 1. Solubility product is written in this case as Kps because it differs
substantially from tabulated Kps values.
1.E-03
8.E-04
Kps'
6.E-04
4.E-04
y = 0.0029x + 0.0001
2.E-04
R = 0.9973
0.E+00
0
0.05
0.1
0.15
0.2
0.25
0.3
m/V
Kps dependence with the temperature was assumed to be the same at all m/V ratios, and
was considered, as described elsewhere19, as:
K ps = 2 10 6 T + 0.0001
Eq. 3
So the higher the temperature is, the lower the solubility product is.
By taking into account this equation, the Kps dependence with the m/V at 333K is
shown in Figure 2:
4.E-04
3.E-04
Kps'
3.E-04
2.E-04
2.E-04
y = 0.0011x + 4E-05
2
R = 0.9973
1.E-04
5.E-05
0.E+00
0.05
0.1
0.15
0.2
0.25
0.3
m/V
the mixture. As described elsewhere21, the value of the solubility product can change
dramatically depending on the particle size of the target solid. It exists an exponential
dependence between both parameters, so from determinate values of particle size, small
variations can give an important raise in the solubility product. By looking at the
Ca(OH)2 particles present in the mixture suspension, an important particle size
reduction can be noticed from the starting to the ending point (see Figures 3a and 3b).
The size of the Ca(OH)2 particles present at the starting point is reduced due to the ion
exchange process that is taking place in the resin. Ca2+ ions from the solution are taken
by the resin, displacing the solubility product equilibrium. Thus, more Ca2+ and OH- are
released from Ca(OH)2 particles surface, resulting in the mentioned particle size
reduction.
approximations are presented and discussed. First of all, it must be taken into account
the separation factor () for the Ca/M system, where M is the target alkaline metal (Li,
Na, K). This coefficient shows the affinity of the resin toward the different elements
present in the solution, so it must be considered in order to predict the uptake of these
elements from the solution by the resin.
Ca
M =
YCa X M q Ca C M
=
YM X Ca q M CCa
Eq. 4
At the same time, it is also necessary to consider the solubility product of Ca(OH)2,
because this solid is the source of Ca in the system.
2
K 'ps = CCa COH
Eq. 5
Taking into account that the resin is initially in M-form, all the M+ ions found in the
solution come from the ion exchange reaction with Ca2+. Then, it can be considered the
following mass balance:
q Ca m = C M V
Eq. 6
The final goal was to have an expression were CM would depend only on one parameter.
The separation factor () was chosen by first, and in consequence, the following
arrangements were done:
q Ca = C M
m
V
C Ca =
K 'ps
Eq. 7
Eq. 8
2
C OH
q M = q o q Ca
Eq. 9
10
Because of the low solubility of Ca(OH)2, it can be approximated that the hydroxyl ion
concentration at the starting point is almost 0. In consequence, such concentration is
directly linked to the Ca2+/M+ ion exchange reaction, and its concentration will be
approximately the same as CM.
C OH C M
Eq. 10
With the combination of all formula and approximations, the following expression is
obtained:
Ca
M
V
CM CM
=
K'
ps q C V
2 o M
CM
Eq. 11
The reorganization of the terms of this expression allows to achieve a fourth degree
equation:
'
m
=0
C 4M + K 'ps Ca
Ca
M C M K ps q o
M
((
Eq. 12
Where qo is a constant, and m/V is kept constant for each modelisation. By keeping also
constant the temperature (separation factor and solubility product depend on it), and
considering the Kps value accordingly to the m/V chosen, the concentration of the ion
M+ and its dependence on can be modelled. However, it is necessary to apply the
Newtons iterative method in order to solve the fourth degree equation:
11
x i +1 = x i
f (x )
Eq. 13
f ' (x )
Where f(x) is the equation found, f(x) its first derivative, xi+1 the result for the
correspondent iteration and xi the result from the previous iteration.
By applying the Newtons iterative method, a value of concentration is found for each
introduced. Figure 4 shows a plot corresponding to CM values from sequential (a.
potassium; b. lithium). The plot is given respectively for five and four different m/V
ratios. At the same time, horizontal lines mark experimental concentrations found for
each m/V ratio, while the diagonal line shows the intersection of each CM vs. plot
with the correspondent final CM regarding the m/V value. Both figures were calculated
for a temperature of 333K.
m/V=0.05
m/V=0.10
1,00
m/V=0.20
0,80
CKOH
m/V=0.15
m/V=0.25
0,60
0,40
0,20
0,00
0
400
800
1200
1600
Alfa
12
m/V=0.05
0,80
m/V=0.15
m/V=0.20
0,60
CLiOH
m/V=0.10
0,70
0,50
0,40
0,30
0,20
0,10
0,00
0
400
800
1200
1600
Alfa
From Figure 4 there can be extracted some conclusions regarding the synthesis system:
-
the higher the m/V ratio is, the higher the final product (MOH) concentration
the higher the m/V ratio is, the lower the separation factor. That means also
lower efficiency (the resin retains a higher amount of M+, because it is less
selective toward Ca2+) and lower purity of the final product
the selectivity of the resin found by modelisation fits very well with previous
studies, as: Li+ > Na+ > K+; in consequence, (Ca/M) has the inverse behaviour
Previous studies have demonstrated the different dependence with XCa for different
temperatures and M+ molarity in the aqueous phase. For this kind of studies, it is
currently recognised the difficulty on finding values working at low XCa values,
13
because of the slowness of the process. From the presented modelisation, there can be
obtained values in the described conditions: low Ca content, due to the low solubility,
different temperatures and different M+ final molarity (it has been shown how it
depends on the m/V ratio). In consequence, a good test for the modelisation goodness is
the correlation with experimentally determined curves at the equilibrium (see Figure
5). As it can be noticed, such correlation is very good. The values obtained from
modelisation (surrounded by circles) fit very well with the experimental ones in a curve
defined as:
y = ae bx + cedx + ge hx
Eq. 14
800
20 C, 0.75 e/L
20 C, 0.2 eq/L
60 C, 0.2 eq/L
60 C, 0.75 eq/L
600
400
200
0
0.0
0.2
0.4
0.6
0.8
1.0
XCa
Figure 5. vs. XCa plot for different temperatures and Na+ content, from experimental
and modelled values
14
Therefore, the proposed model can become a useful tool for the determination of at
extreme conditions, always taking into account some system considerations as
temperature and ionic strength of the process.
F=
q Ca
qo
Eq. 15
F can be also modelled by considering as a variable qCa instead of CM in Eq. 11. In this
case, the achieved expression is:
m
q Ca q Ca
V
=
'
2
K ps V
(q o q Ca )
q2 m2
Ca
Eq. 16
V
V
4
'
'
x + K ps
x q o K ps
=0
m
m
Eq. 17
Where x is the incognita qCa. The degree of conversion is obtained when dividing qCa by
qo. In this case, to check the goodness of the model, it is represented F vs m/V (see
Figure 6). Thus, Kps was taken in each case from experimental values depending m/V,
while was determined in each case by extrapolation from Figure 5.
15
0.9
Experimental
0.8
Model
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.05
0.1
0.15
0.2
0.25
0.3
m/V
Again, the goodness of the model is demonstrated in Figure 6, where data from
modelisation correlates very well with experimental values.
16
Temperature (K)
CK(M) experimental
Error (%)
293
0.479
0.415
13
313
0.458
0.426
333
0.478
0.412
14
Experimental and modelled values appear to have slight differences (error between 7
and 14%). However, it is important to notice the similar tendency with the temperature
in both experimental and modelled values. For the final MOH concentration, the
tendency seems to show almost no variation with the increase of the temperature. Such
tendency can be explained in terms of the different behaviour of some parameters
involved in the synthetic process. Thus, the solubility product (Kps), which is
responsible for the amount of Ca2+ in solution, decreases with the temperature.
Therefore, the higher the temperature is, the lower the concentration of Ca2+ in solution
and in consequence, the lesser the possibility to be changed with M+ from the resin
phase. On the other hand, the temperature dependence of is contrary to the one shown
by Kps. In this case, the higher the temperature is, the higher the separation factor, and
in consequence the higher the affinity of the resin toward Ca2+ ion and to release M+
ions to the solution.
Modelisation of the Ca(OH)2 particle size when working with different m/V ratios
In order to model this process, the evolution of the Ca(OH)2 amount during the
synthetic process was followed. That point allowed us to relate the amount of Ca(OH)2
with the size of its particles.
17
Firstly, the total amount of solid Ca(OH)2 was considered as a set of spherical particles
of a definite radius and density. Thus, next expression is obtained:
2 4 / 3ri3 2 4 / 3rf3
3
3
MW
MW
meq i meq f
= ri rf = R
=
Re l
meq i
ri3
2 4 / 3ri3
MW
Eq. 18
Where meq are the milli equivalents of Ca2+ present in the solid phase (i: initial and f:
final), ri the radius of Ca(OH)2 at the initial solution, rf the radius at the end, the
density of these particles and MW the molecular weight of Ca(OH)2. However, equation
18 can be simplified by considering that the number of particles during the process does
not vary, but the Ca2+ and OH- released to the media come from a homogeneous
dissolution from the surface of all Ca(OH)2 particles. Thus:
R Re l =
ri3 rf3
ri3
Eq. 19
meq i =
C s grCa (OH )2
1mol 2000meq 2V C s 10
VmlH 2 O
=
100mlH 2 O
MW
1mol
MW
Eq. 20
18
Eq. 21
meqf = YCa q o m
Where YCa is the fraction of active sites in the resin occupied by Ca2+ ions. Now, it is
possible to achieve an expression that relates RRel with the ratio m/V.
R Re l
2V C s 10 2V C s 10
YCa q o m
meqi meqf
MW
MW
= YCa q o m MW
=
=
meqi
20V C s
2V C s 10
MW
Eq. 22
However, as RRel does not have a clear physical meaning, it might be substituted by rf/ri
as follows:
rf
= 3 1 R Re l
Eq. 21
ri
Then it can be plotted rf/ri vs. m/V, taking into account that the model must be restricted
by limiting the RRel value, as: 0 < RRel < 1. Figure 7 shows the modelled response.
1.0000
5% Ca(OH)2
0.9000
10% Ca(OH)2
0.8000
15% Ca(OH)2
rf/ri
0.7000
0.6000
0.5000
0.4000
0.3000
0.2000
0.1000
0.0000
0.2
0.4
0.6
0.8
m/V
19
In the Figure 7, the intersection of the different curves with the x-axis shows the
appropriate m/V ratios for the initial Ca(OH)2 concentration used. Then, for an m/V
ratio of 0.23, the most appropriate Ca(OH)2 concentration would be 5%. This estimation
is based on the dependence of the Kps with the particle size, and the implications that
this fact has. Thus, an rf close to 0 means:
-
higher M+ changed for Ca2+ from the aqueous media. Higher MOH
concentration
lower Ca(OH)2 solid present in the synthetic mixture. Less difficulties for phase
separation, and higher purity of the final product.
In consequence, the presented modelisation for rf/ri dependence allows the optimisation
of the Ca(OH)2 concentration, in terms of final product (MOH) concentration and
purity.
Conclusions
This work has demonstrated the strong dependence of the m/V ratio on the solubility
product of Ca(OH)2, for the synthetic process of alkaline hydroxides proposed. This
dependence has been explained by the reduction on the Ca(OH)2 particle size during the
process and depending on the m/V ratio.
The demonstrated dependence of the Kps has been also used in the performed
modelisation studies. These studies have included the modelisation of the final MOH
concentration depending on the separation factor and temperature. It has been also
modelled the variation of the degree of conversion of the resin (F) considering the
variation of the m/V ratio. In all cases, the modelled results have been successfully
20
Acknowledgements
This work has been performed under the economical support of the Ministerio de
Educacin y Ciencia Espaol (project QUI96-1025-C03-01).
Nomenclature
CM: concentration of the M ion in the solution phase, mol/dm3
f(x): original equation in the Newtons iterative method
f(x): first derivative of the original equation in the Newtons iterative method
F: degree of conversion of the resin
Kps(M): Solubility product for the M compound. Tabulated value
Kps(M): Solubility product for the M compound under given conditions of m/V
m: mass of the resin used for each experiment
meq: milli equivalent of a given analyte. i. at the start of the process. f. at the end of the
process
21
Literature cited
(1) Bchner, W. Industrial Inorganic Chemistry, Weinheim: VCH, 1989
(2) PPG company, [on line], <http://www.ppg.com/chm_chloralk/chlor_home.htm>,
[26/12/2002]
(3) Dow chemicals company, [on line],
<http://www.dow.com/causticsoda/product/prdinfo.htm>, [26/12/2002]
(4) Vian Ortuo, A. Introduccin a la Qumica Industrial, Reverte, 1992 (in spanish)
(5) Solvay company, [on line], <http://www.solvay.com>, [26/12/2002]
22
(6) Arribre, M.A. et al. Heavy metals in the vicinity of a chlor-alkali factory in the
upper Negro River ecosystem, Northern Patagonia, Argentina. The Science of the Total
Environment 2003, 301, 187
(7) Masuda, F.Y. Development of environmental pollution protective technology for
mercury-process chlor-alkali cells. Chemical Age of India 1980, 31, 489
(8) Boening, D.W. Ecological effects, transport, and fate of mercury: a general review.
Chemosphere 2000, 40, 1335
(9) Inorganic Mercury. Environmental Health Criteria 118, World Health Organization
(Geneva), 1991
(10) Mercury Environmental Aspects. Environmental Health Criteria 86, World
Health Organization (Geneva), 1989
(11) Mercure. Critres dhygine de lenvironnement 1, Organisation mondiale de la
Sant (Genve), 1977 (in french)
(12) Chrysotile Asbestos. Environmental Health Criteria 203, World Health
Organization (Geneva), 1998
(13) Masuda, F.Y. Recent developments of the chlorine industry in Japan. J. Appl.
Electrochem. 1986, 16, 317
(14) Bolton, H.R. Use of bipolar membranes for ion exchange resin regenerant
production. J. Chem. Tech. Biotech. 1992, 54, 341
(15) Gilbert, K. Industrial Inorganic Chemicals: Production and Uses, The Royal
Society of Chemistry, 1995
(16) Muraviev, D.; Khamizov, R.; Tikhonov, N.A.; Krachak, A.N.; Zhiguleva, T.I.;
Fokina, O.V.
23
(17) Khamizov, R.; Muraviev, D.; Tikhonov, N.A.; Krachak, A.N.; Zhiguleva, T.I.;
Fokina, O.V.
Clean Ion-Exchange
24
Alternative
Ion
Exchange
technologies:
synthesis
of
Sodium
Abstract: This paper deals with the resin regeneration step in an ion exchange
technology for the ecologically synthesis of sodium hydroxide from sodium
chloride and calcium hydroxide by using a weak carboxylic cation exchanger. Two
different approaches for the resin regeneration have been attempted, in order to
complement previous synthesis experiments. On the one hand, a NaCl-rich waste
product from a desalination plant in Mazarrn (in the southern part of Spain) has
been tested without success, despite showing in first term promising results. On
the other hand, a more expensive but effective regeneration process, based on a
NaCl-Na2SO4 mixture as a regeneration solution, has been attempted achieving a
good regeneration rates.
Taking into account these results, an scheme of the complete process is presented,
including both synthesis and regeneration steps.
INTRODUCTION
Typical methodologies for sodium hydroxide production appear to be of a high
efficiency, although they present some environmental troubles. Despite continuous
improvements in the field, chloro-alkali electrolytic plants continue releasing mercury
to the environment1-4, while some other plants still use asbestos for the diaphragm-based
electrolytic process5,6.
In consequence, new technologies as Ion Exchange (IE) technologies must be taken into
account. IE presents several advantages, as the relatively easy design for resin
application, the high purity of the final products (if an adequate system is planned) and
the environment friendly conditions. However, it exists also an important drawback,
which is the resin regeneration. For this final step, aggressive regeneration solutions
must be used, achieving in other case poor regeneration rates. In order to improve the
regeneration step, some modifications of the classical regeneration procedure have been
proposed, i.e. the use of a dual temperature system7-12, which deals with the variation of
the selective coefficients with the temperature, or the Ion-Exchange-IsothermalSupersaturation
(IXISS)
discovered
by
Muraviev13-14,
which
couples
resin regeneration, while the presence of chloride anions seems to be also necessary for
the miscellar stabilisation of the solid CaSO4 particles. Different Cl-/SO42- ratios and
NaCl sources were tested and will be discussed in the following sections.
EXPERIMENTAL
All reactive of p.a. grade were purchased from Panreac (Spain) and used as received.
The carboxylic resin Lewatit CNP-80 was kindly supplied by Bayer Hispania Industrial
S.A., with a total capacity of 7.5meq/g (milli equivalents per gram).
Regeneration experiments were performed on BioRad (US) Econo-Columns, 1.5cm i.d.
Columns were loaded with 1g of resin in Ca-form, and 50mlitre of regenerant solution
was passed through them at a constant flow of 1mlitre/min.
Two types of regenerant mixtures were used. Both combined different proportions of
NaCl and Na2SO4, although the origin of NaCl was different:
-
Mixture 1: 1eq/litre solution of Na2SO4 (p.a. quality) and NaCl solution from a
desalination plant concentrate (in Mazarrn, Murcia). The metallic content of the
concentrate was:
CNa= 0.74eq/litre
CCa= 0.05eq/litre
CMg= 0.18eq/litre
H2O
Sunlight
CaO
H2O
(40C)
(H<0)
Suspension of
Ca(OH)2(s) at 333K
Ca(OH)2(s) Ca2+ + 2OH-
R-Na
(Kps)
Phase separation
Ca(OH)2(s)(exc)
NaOH (product)
R2Ca
Figure 1. Scheme of the proposed system for IE based sodium hydroxide synthesis
After the regeneration treatment, the resin was flushed with 15mlitre of Milli-Q water
and converted to the H-form with HCl 1.5M in order to recuperate the retained cations.
Then, the acidic solutions were analysed by ICP-OES (ARL model 3410, United States,
which was equipped with minitorch) under the specifications detailed in Table 1.
Table 1. Spectroscopic specifications16 for ICP-OES analysis of target elements
Na
Ca
Mg
589.59
422.67
279.806
0.069
0.010
0.015
Interferences
Te, Ti, V
Fe
%R =
Ca
q mEq strip
100
Eq.1
Where %R is the regeneration average, q the total capacity of the resin expressed as
Ca
milli equivalents of Ca, and mEq strip
milli equivalents of Ca after the regeneration
The mixture of Mazarrn water/Na2SO4 (Mixture 1) was firstly tested as the regenerant
solution. The regeneration obtained, calculated as described in Equation 1, is shown in
Figure 2.
29
28
%Regeneration
27
26
25
24
23
22
21
20
0
13
25
37
47
67
%Na2SO4
Figure 2. Regeneration rates for different Na2SO4 content in the regenerant solution. Mazarrn water used as
source of Na+
These results seem to correlate very well with the expected optimum proportion for the
regenerant solution. As described elsewhere17 for the K/Ca system, 25% content of
Na2SO4 seemed to have the appropriate conditions for the regeneration by the IXISS
technology. Although the maximum found at 25% content of Na2SO4 appears to have a
low total regeneration value, it can be improved by flushing the resin with larger
amounts of regenerant solution.
These were the results presented at IEX2000, regarding the last step of the NaOH
synthetic process. However, deeper studies were performed later, related to the
regeneration with Mazarrn water. The evolution of the regenerant solution (Figure 3)
and the metallic content inside the resin after the regeneration treatment (Figure 4) were
analysed in more detail, considering the cations Mg2+ and Ca2+ also present in Mazarrn
concentrate.
0.9
0.16
0.8
0.14
0.7
0.12
0.6
0.1
0.5
0.08
0.4
0.06
0.3
0.04
0.2
0.02
0.1
C(N) (Na)
C(N) (Ca-Mg)
0.18
0
0
10
20
30
40
50
%Na2SO4
Figure 3. Post-column metallic content for the regenerant mixture (Mazarrn water/Na2SO4). : ,
continuous line, left axis: Ca content; , dashed line, left axis: Mg content; , dotted line, right axis: Na
content
6
5
Meq
4
3
2
1
0
0
10
20
30
40
50
%Na2SO4
Figure 4. Metallic content in the resin after regeneration treatment with Mazarrn water/Na2SO4 mixture: ,
continuous line: Ca content; , dashed line: Mg content; , dotted line: Na content
Figure 3 shows only a small drop on the Na+ content in the regenerant solution (in
comparison with the initial 0.74N sodium content), while an important decrease on the
Mg2+ content occurs. The expected result would be the decreasing of Na+ concentration,
because the interchange with the calcium present in the resin and, in consequence, an
increase on the Ca2+ concentration present in the eluate. Correlating with that, an uptake
of Mg2+ by the resin is observed when performing the stripping of the treated resin
(Figure 4), while just a slight increase on the Na+ presence in the resin is observed.
In consequence, no adequate regeneration has been achieved. The presence of Mg2+ in
the regeneration media, despite being in a low content, interferes the equilibrium
between Na+ and Ca2+ and occupies several functional positions in the weak acidic
resin. Because of that, Mazarrn water cannot be used as NaCl source and the
conclusions on the regeneration part presented at the IEX2000 conference must be
discarded.
After the experience with Mazarrn water, pure NaCl solutions combined with Na2SO4
were used in the regeneration step. Regeneration results can be seen in Figure 5.
50
48
46
% Regeneration
44
42
40
38
36
34
32
30
0
10
20
35
50
65
80
100
%Na2SO4
Figure 5. Regeneration rates for different Na2SO4 content in the regenerant solution. Pure NaCl solutions
used as source of Na+
The results obtained for the system pure NaCl/Na2SO4 differ substantially than the
correspondents for Mazarrn water/Na2SO4 and presented at the IEX2000. For the
Na/Ca pair, the behaviour of IXISS process appears to have a different tendency than
the observed for the K/Ca system, having a maximum in the regeneration efficiency for
an 80-100% Na2SO4 content.
CONCLUSIONS
It can be concluded that pure NaCl solutions must be used to prepare the regenerant
mixture with Na2SO4. If waste solutions from desalination plants are used as NaCl
sources, inefficient regeneration is achieved. In that case, the presence of Mg2+ and
Ca2+, despite being in small proportions, interferes dramatically to the regeneration
process because of their higher affinity towards resin functional groups.
When using pure mixtures of NaCl and Na2SO4, different behaviour is obtained.
Differing from previous studies with the K/Ca system, the IXISS effect seems to be
weaker, and the maximum regeneration rates are achieved at high Na2SO4 contents. In
addition, regeneration appears to be much higher than in previous studies17, leading to a
system with lesser consume on regeneration solution. On the other hand, this fact is
consistent with the resin selectivity, since the resin utilised (Lewatit CNP80) appears to
have a higher selectivity toward K+ than Na+.
ACKNOLEDGEMENTS
This work has been performed under the economical support of the Ministerio de
Educacin y Ciencia Espaol (project QUI96-1025-C03-01).
REFERENCES
1. Mishra BB, Kar M, Patro BB, Behera L and Sahoo P, Toxicity degradation of chloroalkali waste soil by Azolla. Indian Journal of Plant Physiology 4:340-342 (1999)
2. Ferreira LM and De Carvalho JMR, Mercury removal from chloro-alkali plant waste
waters by ion exchange. Environ Technol 18:433-440 (1997)
3. Bertani R, Cosimi C, De Liso A, Ferrara R, Maserti BE, Trifoglio M and Zuccarelli
D, Mercury in a primary consumer (Eobania vermiculata) collected near a chlor-alkali
complex. Environ Technol 15:1095-1100 (1994)
4. Yang GCC, Durability study of a solidified mercury-containing sludge. J Hazard
Mater 34:217-223 (1993)
10
16. Winge RK, Fassel VA, Peterson VJ and Floyd MA in Inductively Coupled PlasmaAtomic Emission Spectroscopy. An Atlas of Spectral Information, Elsevier, Amsterdam
(1985)
17. Muraviev D, Noguerol J, Gaona J and Valiente M, Clean Ion-Exchange
technologies. 3. Temperature-Enhanced conversion of Potassium Chloride and lime
milk into Potassium Hydroxide on a carboxylic ion exchange. Ind Eng Chem Res
38:4409-4416 (1999)
11
4409
This paper reports the results obtained by studying the ion-exchange synthesis of potassium
hydroxide from lime milk and carboxylic resin Lewatit CNP 80 in the K-form. Carboxylic resin
has been shown to have several advantages in comparison with, for example, sulfonate ion
exchanger, due to its higher selectivity toward Ca2+, which substantially enhances at elevated
temperature. The increase of temperature from 293 to 333 K enhances the sorption of Ca2+
versus K+ that allows achieving far higher KOH concentration in the solution phase. The values
of equilibrium separation factor, R, for Ca2+-K+ exchange have been determined at 293 and
333 K by varying the equivalent fraction of Ca2+ in the solution phase from 0.04 to 0.7. A
remarkable increase of R values has been observed at higher temperatures and at low calcium
content in the solution phase. The regeneration of the resin (conversion from Ca- to K-form)
has been carried by using the mixtures of potassium chloride and potassium sulfate. The
regeneration process under these conditions is accompanied by the ion-exchange isothermal
supersaturation of calcium sulfate, which forms a stable supersaturated solution in the resin
bed. After leaving the column, CaSO4 crystallizes spontaneously, which allows reuse of the
regenerating solution following the complete elimination of the Ca2+ admixture with a small
amount of K2CO3. The precipitates of calcium sulfate and calcium carbonate are the only wastes
produced in the process. The complete regeneration of the resin has been shown to require a
nearly 70-fold excess of the regenerant; nevertheless, the decrease in the degree of resin
regeneration from 100 to 80% allows reduction of the amount of the regenerant by more than 2
times. The use of incompletely regenerated resin for the synthesis of KOH does not decrease
remarkably the efficiency of the process. The flow sheet of the proposed process is presented
and discussed.
Introduction
Water softening and demineralization processes (Calmon, 1986; Dickenson, 1992; Arden, 1991; Streat, 1991)
still remain the only large-scale industrial applications
of ion exchange (IE). The separation and purification
of high-cost pharmaceuticals (Streat, 1991; Pirotta,
1991), recovery of precious metals such as gold and
platinum from the leaching solutions (Fravel, 1991;
Warshawsky, 1987), and removal of heavy metals from
acidic mine waters are examples of other (usually of far
less scale) applications of IE. An important application
of the IE technique, which still requires further development and wider application, is the IE synthesis of
chemicals.
IE synthesis offers several advantages in comparison
with conventional synthetic procedures such as, ease of
apparatus design, sufficiently high and easily regulated
reaction rates, high purity of the products, and low
consumption of reagents and energy (Vulikh, 1973;
Seniavin, 1980; Dorfner, 1991). At the same time, the
main drawback of the majority of known IE technologies
is the large volume of aggressive liquid wastes (e.g.,
On sabbatical leave from: Department of Physical Chemistry, Lomonosov Moscow State University, 119899 Moscow,
Russia.
Present address: Dpto. de Medio Ambiente, Laboratorio
de Analisis Dr. Echevarne, Provenza 312, Barcelona, Spain.
acidic and/or alkaline) produced during the ion-exchanger regeneration stage. Hence, the enhancement
of this particular stage must lead to significant improvement of the whole process. A possible solution to
this problem is the tailored application of the ionexchange isothermal supersaturation (IXISS) effect,
discovered by Muraviev (Muraviev, 1979; Muraviev and
Saurin, 1980; Muraviev and Fesenko, 1982; Muraviev
and Gorshkov, 1982), for the conversion of ion exchanger
into the desired ionic form, i.e., for regeneration. Use
of the IXISS phenomenon to shift the equilibrium in IE
systems to the desired direction has been shown to allow
for designing highly efficient and ecologically clean IE
technologies (Khamizov et al., 1995, 1996; Gorshkov et
al., 1998; Muraviev et al., 1998a-c).
Another route to increase the efficiency of IE processes is the modulation of temperature in the system,
which in certain instances strongly influences both the
selectivity of the resin and the kinetics of the IE reaction
(Ivanov et al., 1990, 1992; Timofeevskaja et al., 1988;
Muraviev et al., 1995, 1996). The use of the dualtemperature (DT) mode of operation also allows us to
design practically reagentless (and wasteless, as the
result) IE separation processes (Muraviev et al., 1996,
1997a,b; 1999).
During the past decades, the flow sheets of the
majority of IE treatment processes were based on the
fixed-bed mode of operation. In certain cases, this
4410
significantly limits their practical and scaled-up applications. At present, modern IE technology offers a
much wider spectrum of operation modes including fixed
beds, moving packed or fluidized beds, stirred tanks,
and some others (Streat, 1991). The choice of the
optimal process design is mainly determined by the
system parameters such as the concentration and the
amount of suspended solids in the feed solution. For
example, fixed-bed columns are rapidly blocked if the
solution under treatment contains even trace amounts
of suspended solids, while the fluidized bed can be
successfully applied for processing of solutions with fine
solids content up to several hundred grams per cubic
meter. Stirred tanks can be used for the treatment of
slurries containing up to 20 wt % solids (Streat, 1991).
The first two papers of this series reported the results
obtained by studying the IE synthesis of chlorine-free
potassium fertilizers (Muraviev et al., 1998a) and the
high-purity magnesium compounds by using seawater
as the initial magnesium source (Khamizov et al., 1998).
The present (third) paper is dedicated to the IE synthesis of potassium hydroxide. Potassium hydroxide was
originally produced by the reaction between lime and
potassium chloride similar to the soda-lime process used
to manufacture sodium hydroxide. Potassium hydroxide
is now produced exclusively by electrolysis of potassium
chloride brine, in a process analogous to the production
of chlorine and caustic soda. Both diaphragm and
mercury cell processes are used for this purpose (Snell
and Ettre, 1973; Stocchi, 1990; Grayson, 1978). In the
former process, the final product liquor contains 1015% KOH and 10% KCl. Most of the KCl crystallizes
out during the concentration (evaporation) and cooling
stages, but the residual KCl content in the product
(KOH) is usually not less than 1 mass %. In the later
process, the potassium amalgam is separated from the
initial electrolyte (KCl) prior to reacting with water to
produce KOH. As the result, the product obtained is
much purer (contains less than 0.01% KCl) and sufficiently concentrated to be directly marketable (contains 45% KOH). Nevertheless, this advantage of the
mercury cell process cannot compensate for its main
disadvantages dealing with contamination of waste
discharges by mercury.
An attempt to synthesize potassium hydroxide using
a strong acid cation-exchange resin in the K-form and
lime milk or ammonia liquor has been reported in the
literature (Baxi and Krishnaswamy, 1971). Nevertheless, this work has not found any further development
due to the disadvantages of the standard IE processes
mentioned above.
The aim of the present work was to demonstrate the
possibility of design of an economically reasonable and
ecologically clean technology for the synthesis of potassium hydroxide from potassium chloride and lime milk
by applying DT-IXISS-based IE process. This study was
addressed for the following reasons: (1) to select an
appropriate IE resin to carry out the synthesis of
potassium hydroxide; (2) to study the equilibrium of
Ca2+-K+ exchange in systems involving selected ion
exchanger and solution of Ca2+-K+ mixtures of different
compositions; (3) to study the kinetics of Ca2+-K+
exchange on selected resin from systems involving
suspensions of Ca(OH)2 of different concentrations in a
wide temperature range; and (4) to study the regeneration conditions of the resin in the Ca-form by using
IXISS active stripping agents.
Experimental Section
Materials, Apparatus, and Analytical Methods.
The ion exchangers, a sulfonic resin Lewatit SP 112 and
a polyacrylic resin Lewatit CNP 80 bearing a carboxylic
group (both of macroporous type), were kindly supplied
by Bayer Hispania Industrial, S.A. The IE capacities of
the resins equaled 7.5 (Lewatit CNP 80) and 4.5
(Lewatit SP 112) mequiv/g. Potassium chloride, potassium sulfate, and potassium hydroxide of p.a. grade
were purchased from Panreac S.A. Calcium chloride,
sodium hydroxide, and hydrochloric acid of A.G. quality
(Probus S.A.) were used as received. The concentration
of metal ions was determined by atomic emission
spectroscopy (ICP-AES technique) using an ARL model
3410 spectrometer (Fisons) provided with a minitorch.
The uncertainty of the metal ion determination was less
than 1.5%. Determination of H+ and OH- ions was
carried out by potentiometric titration using a Crison
pH-meter 507 provided with a combined glass electrode.
Thermostatic glass columns (of 1.4-cm i.d.) connected
with a thermostat (Haake D1) were used to study the
IE equilibrium at different temperatures. The construction of these columns provided simultaneously the
heating (or cooling) of both the resin and the entering
solution phases. The IE synthesis of potassium hydroxide at different temperatures was carried out in a jacked
cell connected with a thermostat (Grant W6). Regeneration of the resin was performed under dynamic conditions in glass columns of 1.5 20 cm (Bio-Rad).
Methods. Ion-Exchange Synthesis. Experiments
on the IE synthesis of potassium hydroxide were carried
out under batch conditions by using the following
procedure: a portion of the resin in the K-form (1 or 5
g of either sulfonic or acrylic resin) was introduced into
the thermostatic cell preequilibrated at a selected
temperature followed by addition of 20 mL of either 1
or 5% Ca(OH)2 suspension under vigorous stirring. The
moment of lime addition was counted as zero time. After
a certain period of time the stirring was stopped and a
0.1-mL sample of suspension-free supernatant was
withdrawn from the reactor and diluted with 0.1 M HCl
followed by determination of Ca2+ and K+ concentration
in the samples. Experiments on the IE synthesis of KOH
using incompletely regenerated resin were performed
with the mixtures of acrylic resin samples in the Caand K-forms of the desired composition. The total
amount of the resin in this series of experiments in all
cases was constant and corresponded to 5 g.
Ion-Exchange Equilibrium. The IE equilibrium
was studied under dynamic conditions in the thermostatic column. The columns were loaded with a 1-g
portion of Lewatit CNP 80 resin, which remained
constant during all series of experiments. The resin in
the initial Na-form was equilibrated at a selected
temperature with the mixtures of calcium and potassium chlorides of a given composition. The main parameters of the systems studied within this series of
experiments are presented in Table 1.
The attainment of IE equilibrium in the systems
under study (see Table 1) was followed by a comparison
of the metal ion concentrations in the solution leaving
the column with that of the feed. Once the concentration
of Ca2+ and K+ in the eluate was close to that in the
feed solution the flow was stopped and then resumed
after a certain period. The equality of the feed concentration with that of the solution sample collected after
the break was considered as a criterion of achieving the
Ind. Eng. Chem. Res., Vol. 38, No. 11, 1999 4411
Table 1. System Parameters and Results Obtained by
Studying K+-Ca2+ Exchange Equilibrium on Lewatit
CNP 80 Resin at 293 and 333 K
T, K [Ca], equiv/L [K], equiv/L
293
293
293
293
333
333
333
333
293
293
293
293
333
333
333
333
0.045
0.135
0.237
0.320
0.031
0.127
0.236
0.321
0.006
0.015
0.026
0.045
0.006
0.015
0.026
0.037
0.639
0.657
0.529
0.401
0.639
0.633
0.530
0.401
0.077
0.073
0.059
0.045
0.077
0.073
0.059
0.046
XCa
Ctot, equiv/L
YCa
0.065
0.171
0.310
0.444
0.047
0.167
0.308
0.444
0.077
0.169
0.302
0.476
0.077
0.169
0.302
0.444
0.7
0.8
0.8
0.7
0.7
0.8
0.8
0.7
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.81
0.88
0.92
0.95
0.87
0.96
0.97
0.98
0.91
0.94
0.96
0.97
0.97
0.97
0.98
0.99
60.3
36.4
27.2
22.2
135
114
61.2
50.9
129
75.5
51.1
26.3
328
183
126
101
RCa
K )
(1)
4412
(2)
RK
Ca )
(3)
Ind. Eng. Chem. Res., Vol. 38, No. 11, 1999 4413
LCaSO4 ) CCaCSO42-
(4)
(5)
RK
Ca )
LCaSO4
(1 - YCa)
YCa
Co(Co - CS)
(6)
containing 75% KCl and 25% K2SO4 that can be attributed to the maximum stability of supersaturated
solution of calcium sulfate formed in the interstitial
space of the column when the IXISS-active stripping
agent of this particular composition is applied (Muraviev
et al., 1997c, 1998c). The stabilization mechanism in this
case is associated with the sorption of Cl- ions on the
surface of precrystalline CaSO4 aggregates (micelles).
The resulting negative charges arising on the micelle
surfaces prevent their further aggregation and make
impossible their sorption on the functional groups of the
ion exchanger (also bearing negative charges). The resin
phase in this case serves as an additional stabilization
factor toward supersaturated solution. Hence, the stabilization of CaSO4 supersaturated solution in the
interstitial space of the column appears to be due to a
double-function mechanism (Muraviev et al., 1997c).
The substantiation of this hypothesis is provided by the
results obtained by using stripping solutions with higher
(>25%) K2SO4 content. For example, elution of the resin
with pure K2SO4 solution resulted in the fast formation
of CaSO4 precipitate in both the column and the eluate.
The decrease of K2SO4 content in the regenerating
solution (due to addition of KCl) led to a gradual
decrease of the amount of precipitate up to its complete
disappearance at 25% K2SO4 in the KCl-K2SO4 mixture. This indicated the formation of a stable supersaturated solution of CaSO4 in the interstitial space of
the column. Moreover, the eluate collected in this
experiment started to crystallize after 1 h following
its removal from the column.
The typical concentration-volume history obtained
by regeneration of Lewatit CNP 80 in the Ca-form with
75% KCl-25% K2SO4 solution at 293 K is shown in
Figure 5a. As seen, the breakthrough curve of Ca2+
displacement is characterized by a very long tail associated with a slow kinetics of Ca2+-K+ exchange in the
system. This results in the need to pass nearly a 70fold excess (in equivalents per 1 equiv of the resin
functional groups) of the regenerant through the column
to achieve the complete removal of Ca2+ from the resin
(100% regeneration). On the other hand, as follows from
the results presented in Figure 5b, incomplete regeneration of the resin (for example up to 80% instead of
4414
Ind. Eng. Chem. Res., Vol. 38, No. 11, 1999 4415
Literature Cited
(1) Arden, T. V. Raw Water Treatment by Ion Exchange. In
Ion Exchangers; Dorfner, K., Ed.; Walter de Gruyter: Berlin, 1991;
pp 717-789.
(2) Baxi, D. R.; Krishnaswamy, N. Causticization of Alkali
Metal Ion Fixed on a Cation Exchange Resin Using Lime and
Aqueous Ammonia. Indian J. Technol. 1971, 9 (8), 300.
(3) Calmon, C. Recent Developments in Water Treatment by
Ion Exchange. React. Polym, 1986, 4, 131.
(4) Dickenson, Ch. Filters and Filtration Handbook, 3rd ed.;
Elsevier Advanced Technology: Oxford, 1992; p 358.
(5) Dorfner, K. Ion Exchangers as Preparative Agents. In Ion
Exchangers; Dorfner, K., Ed.; Walter de Gruyter: Berlin, 1991;
pp 1119-1160.
(6) Fravel, H. G. Electroplating Industry and Metal Recovery.
In Ion Exchangers; Dorfner, K., Ed.; Walter de Gruyter: Berlin,
1991; pp 903-919.
4416